Glossary

8-bit character

A single-byte character that consists of eight bits, which means that the code point is in the range 128 through 255. Examples from the ISO8859-1 code set include the non-English é, ñ, and ö characters. They can be interpreted correctly only by software that is 8-bit clean. See also non-ASCII character.

8-bit clean

An attribute of software that can process character data that contains 8-bit characters. The operating system or the database server reads the eighth bit as part of the code value. In other words, it does not ignore the eighth bit nor make its own interpretation of the eighth bit.

16-bit code set

Any code set that represents each character by two bytes, so that approximately 65,000 distinct characters can be encoded. Also called double-byte code set. An example is JIS X0208.

access method

Any of a group of routines that access or manipulate a table or an index. In the output of a SET EXPLAIN statement, access method refers to the mode of table access in a query (for example, SEQUENTIAL SCAN as opposed to INDEX PATH). See also primary access method and secondary access method.

access privileges

The types of operations that a user has permission to perform in a specific database or database object. The Informix database, table, table-fragment, index, and column-access privileges are independent of operating-system access permissions.

active set

The collection of rows that satisfies a query associated with a cursor.

aggregate function

An SQL function that returns one value for a group of retrieved rows; for example, the frequency, sum, average, maximum, or minimum of an expression in a query or report. See also user-defined aggregate.

aggregate support function

One of a group of user-defined functions that the database server uses to calculate a user-defined aggregate.

alias

In an SQL query or in a form-specification file, an alias is a single-word temporary alternative name that is used in place of a qualified table name (for example, t1 as an alias for owner.table_name). Aliases are often used in complex subqueries and are required for a self-join.

ALS

Legacy acronym for Asian Language Support (ALS), a feature for multibyte East Asian locales. Supplanted by Global Language Support (GLS).

ANSI

Acronym for the American National Standards Institute. This group sets standards in many areas, including the information technology industry and the SQL language. Some ANSI standards are also standards of the International Standards Organization (ISO).

ANSI compliant

For databases, conforming to ANSI/ISO standards for the SQL language. An IBM Informix database can be created as either ANSI compliant or not ANSI compliant. An ANSI-compliant database enforces certain ANSI requirements, such as implicit transactions, explicit owner naming, no public synonyms, and unbuffered logging, which are not enforced in databases that are not ANSI compliant.

API

See application programming interface (API).

application development tool

Software, such as IBM Informix 4GL, which you can use to create and maintain a database. The software allows a user to send instructions and data to (and receive information from) the database server.

application process

The process that manages an ESQL or other program at runtime. It executes the program logic and initiates SQL requests. Memory that is allocated for program variables, program data, and the fetch buffer is part of this process. See also database server process.

application-productivity tools

Software tools that are used to write application programs.

application program

An executable file or a logically related set of files.

application programming interface (API)

A set of related software components, such as those provided by IBM, that a developer uses to create applications that communicate with a third-party product. An API can include a library of functions, header files, graphic interfaces, and application programs. See also SQL API and DataBlade API.

arbitrary rule

Expressions that define a fragmentation strategy. Unlike a range rule, an arbitrary rule can specify any relational and logical operators of SQL to define expressions (such as the OR operator to group data).

archiving

To copy all the data and indexes of a database to a different physical device from that which stores the database. Archived material can be used by the Database Administrator for recovery from a failure. See also backup.

argument

A value that is passed to a routine or command. Compare with parameter.

array

An ordered set of items of the same data type. An individual member of the array is referred to as an element and usually is identified by an integer index that gives the position of the element within the array.

ASCII

Acronym for American Standards Committee for Information Interchange, a code scet of letters, digits, punctuation, and certain other nonprintable and printable characters used in computers and telecommunication. It contains 128 characters, each of which can be represented with 7 bits of information. The code set of every Informix locale includes the ASCII characters as a subset. See also single-byte character.

ASF

Acronym for Associated Services Facility. The ASF code in IBM Informix products controls the connections between client applications and database servers. Software architects use this term; most users of IBM Informix products see this term only in occasional error messages.

Asian Language Support (ALS)

A legacy technology to support the multibyte code sets of certain East Asian languages. The functionality of ALS is replaced by Global Language Support (GLS) in current IBM Informix products. For more information about ALS and GLS, see the IBM Informix: Migration Guide.

attached index

An index that has the same distribution scheme as its table. To create an attached index, specify no distribution scheme (no FRAGMENT BY clause) and no storage option (no IN clause), or else specify IN TABLE as the storage option in CREATE INDEX or ALTER FRAGMENT ON INDEX statements. For Dynamic Server, index pages for user indexes reside in separate tblspaces, but in the same dbspaces as the data pages to which they refer. Only system indexes reside in the same tblspace as the corresponding data pages. For Extended Parallel Server, user index and system index pages reside in separate tblspaces, but in the same dbspaces as the corresponding data pages. See also detached index, system index, user index.

audit event

Any database server activity or operation that can potentially access or alter data. Audit events can be recorded and monitored by the database secure auditing facility. Examples of audit events include accessing tables, altering indexes, dropping chunks, granting database access, updating the current row, running database utilities, and so forth. (For a complete list of audit events, see the IBM Informix: Trusted Facility Guide.)

audit file

A file that contains records of audit events and resides in the specified audit directory. Audit files provide an audit trail of information that can be extracted by the database secure auditing facility for analysis by the database system security officer (DSSO).

audit mask

A structure that specifies which audit events should be logged (or excluded from logging) by the database secure auditing facility.

auxiliary statements

A residual category of SQL statements, including statements like INFO, OUTPUT, and GET DIAGNOSTICS that can display information about the database.

B-tree

A model (with "simple graph" topology) for defining the logical structure of an index.

B-tree index

A type of index that uses a balanced tree structure for efficient record retrieval. A B-tree index is balanced when the leaf nodes are all at the same level from the root node. B-tree indexes store a list of rowids for any duplicate key value data in ascending or descending order. See also bitmap index and R-tree index.

backup

A duplicate of a computer file on another device or tape to preserve existing work, in case of a computer failure or other mishap. A backup refers to duplicating logical-log files, while archiving refers to duplicating data.

base table

(1) A table referenced in the definition of a view, containing data that the view displays. A multitable view has more than one base table.

(2) Within a table hierarchy of Dynamic Server, a parent table from whose schema another table (the child table) is derived. See also table hierarchy.

base type

See opaque data type.

before-image

The image of a row, page, or other item before any changes are made to it.

bidirectional

Of a locale, supporting both right-to-left and left-to-right text.

big-endian

A hardware-determined storage method in which the most-significant byte of a multibyte number has the lowest address. See also little-endian.

bitmap index

A type of index that stores a bitmap for any highly duplicate key value. The bitmap indicates which rows have the duplicate key value. You create a bitmap index with the USING BITMAP keywords in the CREATE INDEX statement. See also B-tree index.

blob

A legacy term for a large object that is now known as a simple large object and includes TEXT or BYTE data types. These data objects effectively have no maximum size (theoretically as large as 2³¹ bytes). See also simple large object.

BLOB

Acronym for binary large object. A data type for a smart large object that stores any type of binary data, including images. These objects can be stored and retrieved in pieces and have database properties such as recovery and transaction rollback. See also CLOB.

blobpage

(Not for Extended Parallel Server) The unit of disk allocation within a blobspace. The database server administrator determines the size of a blobpage. The size can vary, depending on the size of the TEXT or BYTE data that the user inserts.

blobspace

(Not for Extended Parallel Server) A logical collection of chunks that is used to store TEXT and BYTE data.

Boolean

A variable or an expression that can take on the logical values TRUE (1), FALSE (0), or UNKNOWN (if NULL values are involved).

BOOLEAN

A built-in data type that supports single-byte true/false values. TRUE is represented internally as 0 and externally as t. FALSE is represented internally as 1 and externally as f. A NULL value is represented as NULL.

Boolean function

A function that returns a Boolean value (true or false). A Boolean function can act as a filter.

branch node

An index page that contains pointers to a leaf node or other branch nodes. The database server creates a branch node when the root node and subsequent leaf nodes become full.

buffer

A portion of computer memory where a program temporarily stores data. Data typically is read into or written out from buffers to disk.

buffered disk I/O

Disk I/O that is controlled by the operating system instead of by an application. With buffered disk I/O, the operating system stores data in the kernel portion of memory before periodically writing the data to disk. See also unbuffered disk I/O and disk I/O.

buffered logging

A type of logging that holds transactions in a memory buffer until the buffer is full, regardless of when the transaction is committed or rolled back. Informix database servers provide this option to speed up operations by reducing the number of disk writes.

built-in

Provided by the database server, usually in the system catalog; not defined by the user.

built-in data type

A predefined data type that the database server supports; for example, INTEGER, CHAR, or SERIAL.

byte

An physical computer storage unit of 8 bits. A character is not necessarily one byte. In a multibyte code set, a character can require more than one byte.

BYTE

A data type for a simple large object that stores any type of binary data and can be as large as 2³¹ bytes. See also TEXT.

Cartesian product

The set that results when you pair each and every member of one set with each and every member of another set. A Cartesian product results from a multiple-table query when you do not specify the joining conditions among tables. See also join.

cascading deletes

Deletion of rows from a child table that were associated by foreign key to a row that is deleted from the parent table. When any rows are deleted from the primary key column of a table, cascading deletes, if enabled, delete identical information from any foreign-key column in a related table.

case sensitivity

The condition of distinguishing between uppercase and lowercase letters. Case-sensitive commands and command options react differently to the same letters entered as uppercase or lowercase characters. The locale files specify which characters (if any) of a code set are uppercase or lowercase.

cast

A database object that converts one data type to another. Most built-in data types have built-in casts (that is, system-defined casts) to compatible data types. See also user-defined cast, explicit cast, and implicit cast.

cast function

A user-defined function that implements a cast. The function must be registered with the CREATE CAST statement before it can be used.

character

A logical unit of storage for a code point. A character is equal to one or more bytes and can be numeric, alphabetic, or a nonprintable character (such as a control character). See also multibyte character and single-byte character.

character set

One or more natural-language alphabets, together with additional symbols for digits, punctuation, ligatures, diacritical marks, and whitespace. (A natural language is a written language that human beings use to communicate with each other, such as English, Chinese, or German.) See also code set.

character special device

See unbuffered disk I/O.

check constraint

A logical condition that must be satisfied before data values can be entered into a column of a database table during an INSERT or UPDATE statement.

checkpoint

A point during a database server operation when the pages on disk are synchronized with the pages in the shared memory buffer pool. It can be a full checkpoint or a fuzzy checkpoint.

child table

The referencing table in a referential constraint. See also parent table.

chunk

The largest contiguous section of disk space available for a database server. A specified set of chunks defines a dbspace or blobspace. A database server administrator allocates a chunk to a dbspace or blobspace when that dbspace or blobspace approaches full capacity. A chunk contains a certain number of pages. (Extended Parallel Server does not support blobspace chunks.)

client application

A program that requests services from a server program, typically a file server or a database server. For the GLS feature, the term client application includes database server utilities.

client computer

The computer on which a client application runs.

client locale

The locale that a client application uses to perform read and write operations on the client computer. The CLIENT_LOCALE environment variable can specify a nondefault locale. See also server locale.

client/server architecture

A hardware and software design that allows the user interface and database server to reside on separate nodes or platforms on a single computer or over a network. See also ASF, client application, and server-processing locale.

client/server connection statements

The SQL statements that can make connections with databases. These statements include CONNECT, DISCONNECT, and SET CONNECTION.

CLOB

Acronym for character large object. A data type for a smart large object that stores large text items, such as PostScript or HTML files. These objects can be stored and retrieved in pieces and have database properties such as recovery and transaction rollback. See also BLOB.

close a cursor

To drop the association between a cursor and the active set of rows that results from a query.

close a database

To deactivate the connection between a client application and a database. Only one database can be active at a time.

close a file

To deactivate the association between a file and a program.

cluster an index

To rearrange the physical data of a table according to a specific index.

cluster key

The column in a table that logically relates a group of simple large objects or smart large objects that are stored in an optical cluster.

clustersize

The amount of space, specified in kilobytes, that is allocated to an optical cluster on an optical volume.

code point

A bit pattern that represents one character in a code set. For example, in the ASCII code set, the uppercase A character has a code point of 0x41.

code set

The representation of a character set that specifies how to map each element of a character set to a unique code point. For example, ASCII, ISO8859-1, Microsoft 1252, and EBCDIC are code sets to represent the English language. A locale name specifies a code set.

code-set conversion

The process of converting character data from one code set (the source code set) to another (the target code set). Code-set conversion is useful when the client and server computers use different code sets to represent the same character data.

code-set order

The serialized order of characters within a code set. For example, in the ASCII code set, uppercase characters (A through Z) are ordered before lowercase characters (a through z). See also collation order and localized order.

cogroup

A named group of coservers. At initialization, the database server creates a cogroup that is named cogroup_all from all configured coservers.

collating sequence

See collation order, code-set order, and localized order.

collation

The process of sorting character strings according to some order. The term is sometimes used as a synonym for collation order or for localized order.

collation order

The logical order in which the character-string values in a database are sorted and indexed. This is based on either the order of the code set or else some locale-specific order. Collating order is also known as collating sequence.

collection

An instance of acollection data type; a group of elements of the same data type stored in a SET, MULTISET, or LIST.

collection cursor

A database cursor that has an IBM Informix ESQL/C collection variable associated with it and provides access to the individual elements of a column whose data type is a collection data type.

collection data type

A complex data type whose instances are groups of elements of the same data type, which can be any opaque data type, distinct data type, built-in data type, collection data type, or row type.

collection-derived table

A table that IBM Informix ESQL/C or SPL creates for a collection column when it encounters the TABLE keyword in an INSERT, DELETE, UPDATE, or SELECT statement. ESQL/C and SPL store this table in a collection variable to access elements of the collection as rows of the collection-derived table.

collection subquery

A query that takes the result of a subquery and turns it into a expression by using the MULTISET keyword to convert returned values into a MULTISET collection.

collection variable

An IBM Informix ESQL/C host variable or SPL variable that holds an entire collection and provides access, through a collection cursor, to the individual elements of the collection.

collocated join

A join that occurs locally on the coserver where the data resides. The local coserver sends the data to the other coservers after the join is complete.

column

A data element that contains a specified type of information that occurs in every row of the table. Also known as a display label or a field. See also row.

column expression

An expression that includes a column name and optionally uses column subscripts to define a column substring.

column subscript

A subscript (an integer enclosed between brackets) in a column expression, showing the ordinal position of a byte within the column value. Two integers, separated by a comma, as in col2[3,9], define a column substring.

column substring

A substring of a character column. For example, the column expression col2[3,9]specifies a 7-byte substring of col2 that begins with the 3rd byte.

command file

A system file that contains one or more statements or commands, such as SQL statements.

comment

Information in a program file that is not processed by the computer but that documents the program. Special characters such as the sharp sign ( # ), braces ( { } ), slash mark followed by asterisk ( / * ), or a double hyphen ( -- ) can be used to identify comments, depending on the programming context.

commit work

To complete a transaction by accepting all changes to the database since the beginning of the transaction. See also roll back.

Committed Read

An Informix level of isolation in which the user can view only rows that are currently committed at the moment when the query is requested; the user cannot view rows that were changed as a part of a currently uncommitted transaction. Committed Read is available through a database server and set with the SET ISOLATION statement. It is the default level of isolation for databases that are not ANSI compliant. See also Read Committed.

compatible data types

Two data types for which casts exist in the database. See also implicit cast.

compile

To translate source code (in a high-level language) into executable code. Compare with execute and link. See also source file.

compile-time error

An error that occurs when you compile the program source code. This type of error indicates syntax errors in the source code. Compare with runtime error.

complex data type

A data type that is built from a combination of other data types using an SQL type constructor and whose components can be accessed through SQL statements. See also row type and collection data type.

component

In the High-Performance Loader (HPL), the information required to load or unload data is organized in several components. The components are format, map, filter, query, project, device array, load job, and unload job.

composite data type

See row type.

composite index

An index constructed on two or more columns of a table. The ordering imposed by the composite index varies least frequently on the first-named column and most frequently on the last-named column.

composite join

A join between two or more tables based on the relationship among two or more columns in each table. See also simple join.

compressed bitmap

An indexing method that identifies records through a fragment identifier and a record identifier.

concatenate

To append a second string to the end of a first string.

concatenation operator

An operator whose notation is composed of two pipe symbols (||); this is used in expressions to indicate the joining of two strings.

concurrency

Access by two or more processes to the same database simultaneously.

configuration management (CM) coserver

A coserver designated to run CM software and store CM data.

configuration file

A file read during database server disk or shared-memory initialization that contains the parameters that specify values for configurable behavior. Database server and its archiving tool use configuration files.

connection

A logical association between two applications or between an application and a database environment, created by a CONNECT or DATABASE statement. Database servers can also have connections to one another. See also explicit connection, implicit connection, and multiplexed connection.

connection coserver

The coserver to which a client is directly connected. See also coserver and participating coserver.

connection redirector

An Extended Parallel Server feature, enabled by a setting in the sqlhosts file, whereby the database server attempts to establish a client connection with each coserver in a dbserver group until a connection succeeds.

constant

A value that cannot change during the execution of a program or command. In some programming languages, a constant has a name that can be referenced. Compare with variable. See also literal.

constraint

A restriction on what kinds of data can be inserted or updated in tables. See also check constraint, primary-key constraint, referential constraint, not-NULL constraint, and unique constraint.

constructed data type

See complex data type.

constructor

See type constructor.

control character

A character whose occurrence in a particular context initiates, modifies, or stops a control function (an operation to control a device, for example, in moving a visual cursor or in reading data). In a program, you can define actions that use the CTRL key with another key to execute some programming action (for example, entering CTRL-W to obtain online Help in IBM Informix products). A control character is sometimes referred to as a control key. Compare with printable character.

cooked files

See buffered disk I/O.

coordinating server

In a query that spans multiple database servers, the server in which the query is initiated is called the coordinator or coordinating server. This server is also sometimes called the local server because it is the local server to the client initiating the query. To respond to the query, the coordinating server starts sessions on the other servers involved in the query. See also distributed query, subordinate servers, and remote servers.

correlated subquery

A subquery (or inner SELECT) that depends on a value produced by the outer SELECT statement that contains it. Also a nested subquery whose WHERE clause refers to an attribute of a relation that is declared in an outer SELECT. Correlated subqueries reference one or more columns from one or more tables of a parent query and need to be evaluated once for each row in the parent query. See also independent subquery and subquery.

correlation name

The prefix used with a column name in a triggered action to refer to an old (before triggering statement) or a new (after triggering statement) column value. The associated column must be in the triggering table. See also trigger.

corrupted database

A database whose tables or indexes contain incomplete, inconsistent, or invalid data.

corrupted index

An index that does not correspond exactly to the data in its table.

coserver

The functional equivalent of a database server that operates on a single node. See also connection coserver and participating coserver.

current row

The most recently retrieved row of the active set of a query.

cross-server query

See distributed query.

cursor

In SQL, an identifier associated with a group of rows or with a collection data type. Conceptually, the pointer to the current row or collection element. You can use cursors for SELECT statements or EXECUTE PROCEDURE statements (associating the cursor with the rows returned by a query) or INSERT statements (associating the cursor with a buffer to insert multiple rows as a group). A select cursor is declared for sequential only (regular cursor) or nonsequential (scroll cursor) retrieval of rows. In addition, you can declare a select cursor for update (initiating locking control for updated and deleted rows) or WITH HOLD (so that completing the current transaction does not close the cursor). In ESQL/C, a cursor can be dynamic, meaning that it can be referenced by an identifier or by a character variable.

cursor function

A user-defined routine (UDR) that returns one or more rows of data and therefore requires a cursor to execute. An SPL routine is a cursor function when its RETURN statement contains the WITH RESUME keywords. An external function is a cursor function when it is defined as an iterator function. Compare with noncursor function.

cursor manipulation statements

The SQL statements that control cursors; specifically, the CLOSE, DECLARE, FETCH, FLUSH, OPEN, and PUT statements.

Cursor Stability

An Informix level of isolation available through the SET ISOLATION statement in which the database server must secure a shared lock on a fetched row before the row can be viewed. The database server retains the lock until it receives a request to fetch a new row. See also isolation.

data access statements

The subset of SQL statements that you can use to grant and revoke permissions and to lock tables.

data definition statements

The subset of SQL statements (sometimes called data definition language, or DDL) to create, alter, drop, and rename data objects, including databases, tables, views, synonyms, triggers, sequences, and user-defined routines.

data dictionary

The set of tables that keeps track of the structure of the database and the inventory of database objects. This is also called the system catalog. Each database that a database server supports has its own system catalog.

data distribution

A mapping of the data values within a column into a set of categories that are equivalent to a histogram or to a frequency distribution. You can use the UPDATE STATISTICS statement (specifying the MEDIUM or HIGH keyword options) to create data distributions.

data integrity

The process of ensuring that data corruption does not occur when multiple users simultaneously try to alter the same data. Locking, constraints, and transaction logging are used to control data integrity.

data integrity statements

SQL statements that you use to control transactions and audits. Data integrity statements also include statements for repairing and recovering tables.

data manipulation statements

The SQL statements that can query tables, insert into tables, delete from tables, or update tables (select, insert, delete, update). The load and unload utilities also are sometimes called data manipulation statements.

data partitioning

See table fragmentation.

data replication

The ability to allow database objects to have more than one representation at more than one distinct site.

data restriction

Synonym for constraint.

data type

A descriptor assigned to each column in a table or program variable, which indicates the type of data the column or program variable is intended to hold. Informix data types are discussed in Chapter 2, "Data Types." Informix data types for Global Language Support are discussed in the IBM Informix: GLS User's Guide. See also built-in data type, complex data type, distinct data type, opaque data type, and user-defined data type.

database

A collection of information (contained in tables) that is used for a specific purpose. See also relational database.

Database Administrator

See DBA.

database application

A program that applies database management techniques to implement specific data manipulation and reporting tasks.

database environment

Used in the CONNECT statement, either the database server or the database server-and-database to which a user connects.

database locale

The locale that defines the code set, collation, and date, time, number, and currency display conventions of a database server. The DB_LOCALE environment variable can specify this locale.

database management system

See DBMS.

database object

An SQL entity that is recorded in a system catalog table, such as a table, column, constraint, access method, default value, cast, index, operator class, prepared statement, privilege, role, sequence, synonym, trigger, user-defined aggregate, user-defined cast, user-defined data type, user-defined routine, or view.

database server

A software package that manages access to one or more databases for one or more client applications. See also relational database server.

database server process

A virtual processor that functions similarly to a CPU in a computer. See also application process.

database server utility

A program that performs a specific task. For example, DB–Access, dbexport, and onmode are Informix database server utilities.

DataBlade API

An application programming interface (API) that allows a C user-defined routine access to the client application.

DataBlade module

A group of database objects and supporting code that extends an object-relational database to manage new kinds of data or add new features. A DataBlade module can include new data types, routines, casts, aggregates, access methods, SQL code, client code, and installation programs.

DBA

Acronym for Database Administrator. The DBA is responsible for the contents and use of a database, whereas the database server administrator (DBSA) is responsible for managing one or more database servers. Also a level of privilege, typically for operations that most users are not authorized to perform.

DBA-privileged

A class of SPL routines that only a user with DBA database privileges creates.

DBMS

Acronym for database management system. These are all the components necessary to create and maintain a database, including the application development tools and the database server.

dbserver group

A collection of coservers defined and named by entries in the sqlhosts file. Dbserver groups make multiple coservers into a single logical entity for establishing or changing client/server connections.

dbslice

A named set of dbspaces that can span multiple coservers. A dbslice is managed as a single storage object. See also logslice, physslice, and rootslice.

dbspace

A logical collection of one or more chunks. Because chunks represent specific regions of disk space, the creators of databases and tables can control where their data is physically located by placing databases or tables in specific dbspaces. A dbspace provides a link between logical (such as tables) and physical units (such as chunks) of storage. See also root dbspace.

DDL

Acronym for data definition language, a subset of the Structured Query Language (SQL) for declaring, modifying, and dropping database objects (such as tables, constraints, or indexes). See also data definition statements.

deadlock

A situation in which two or more threads cannot proceed because each is waiting for data locked by the other (or another) thread. The database server monitors and prevents potential deadlock situations by sending an error message to the application if a request for a lock might result in a deadlock.

debug file

A file that receives output used for debugging purposes.

decision-support application

An application that provides information that is used for strategic planning, decision-making, and reporting. It typically executes in a batch environment in a sequential scan fashion and returns a large fraction of the rows scanned. Decision-support queries typically scan the entire database. See also online transaction processing application.

decision-support query

A query that a decision-support application generates. It often requires multiple joins, temporary tables, and extensive calculations, and can benefit significantly from PDQ. See also online transaction processing queries.

declaration statement

A programming language statement that describes or defines objects; for example, defining a program variable. Compare with procedure. See also data definition statements.

default

Values or behavior that take effect unless the user explicitly specifies another value or action.

default locale

The locale that a product uses unless you specify a different (nondefault) locale. For IBM Informix products, U.S. English is the default locale.

default value

A value that is used when no explicit value is specified. For example, you can assign default values to columns with the ALTER TABLE and CREATE TABLE statements and to variables in SPL routines.

delete

To remove any row or combination of rows with the DELETE statement.

delimited identifier

If the DELIMIDENT environment variable is set, this is an SQL identifier enclosed between double ( " ) quotation marks. This supports identifiers that are SQL-reserved keywords or that contain whitespace or other characters outside the default SQL character set for identifiers.

delimiter

A character used as a boundary on a field or as the terminator for a column or row. Some files and prepared objects require semicolon ( ; ), comma ( , ), blank space ( ), or tab delimiters between statements. SQL statements can have semicolon or other delimiters in some programming contexts.

deluxe mode

A method of loading or unloading data that uses regular inserts.

descriptor

A quoted string or variable that identifies an allocated system-descriptor area or an sqlda structure. It is used for the Informix SQL APIs. See also identifier.

detached index

An index whose distribution scheme (specified by the FRAGMENT BY clause) and storage option (specified by the IN clause) of the CREATE INDEX or ALTER FRAGMENT ON INDEX statement differ from the distribution scheme of the underlying table. Index pages reside in separate dbspaces from the corresponding data pages. Compare with attached index.

device array

A list of I/O devices. See also component.

diagnostic area

A data structure (sometimes called sqlda) that stores diagnostic information about an executed SQL statement.

diagnostics table

A special table that holds information about the integrity violations caused by each row in a violations table. You use the START VIOLATIONS TABLE statement to create violations tables and diagnostics tables and associate them with a base table.

Dirty Read

An Informix isolation level set with the SET ISOLATION statement that disregards locks and allows viewing of any existing rows, even rows that currently can be altered from inside an uncommitted transaction. Dirty Read is the lowest level of isolation (no isolation at all), and is thus the most efficient. See also Read Uncommitted.

disabled mode

The object mode in which a database object is disabled. When a constraint, index, or trigger is in the disabled mode, the database server acts as if the object does not exist and does not take it into consideration during the execution of data manipulation statements.

disk configuration

The organization of data on a disk; also refers to the process of preparing a disk to store data.

disk I/O

The process of transferring data between memory and disk. The I/O refers to input/output.

display label

A temporary name for a column or expression in a query.

distinct data type

A data type that you declare with the CREATE DISTINCT TYPE statement. A distinct data type has the same internal storage representation as its source type (an existing opaque data type, built-in data type, named row type, or distinct data type) but a different name, and can have different casts and routines. To compare a distinct data type with its source type requires an explicit cast. A distinct data type inherits all routines that are defined on its source type.

distributed query

A query that accesses data from a database other than the current database.

distribution

See data distribution.

distribution scheme

See table fragmentation.

DLL

See dynamic link library (DLL).

DML

Acronym for data manipulation language. See also data manipulation statements.

dominant table

See outer join.

DRDA

Acronym for Distributed Relational Database Architecture. DRDA is an IBM-defined set of protocols that software manufacturers can follow to develop connectivity solutions between heterogeneous relational database management environments.

DSS

Acronym for Decision Support System. See also decision-support application.

duplicate index

An index that allows duplicate values in the indexed column.

dynamic link library (DLL)

A shared-object file on a Windows system. See also shared library.

dynamic management statements

The SQL statements that describe, execute, and prepare other statements.

dynamic routine-name specification

The execution of a user-defined routine whose name is determined at runtime through an SPL variable in the EXECUTE PROCEDURE, EXECUTE ROUTINE, or EXECUTE FUNCTION statement.

Dynamic Server instance

The set of processes, storage spaces, and shared memory that together comprise a complete database server. A single Dynamic Server instance can support more than one database.

dynamic SQL

The statements and structures that allow a program to form an SQL statement at runtime, so that portions of the statement can be determined by user input.

dynamic statements

The SQL statements that are specified at runtime (when the program is executed), rather than when the program is compiled. You can use the PREPARE statement to create dynamic SQL statements.

EBCDIC

Acronym for Extended Binary Coded Decimal Interchange Code, a 256-element 8-bit character set, originally designed for IBM mainframe computers.

element

A member of a collection, such as a LIST, MULTISET, or SET data type. An element can be a value of any built-in data type, opaque data type, distinct data type, named row type, unnamed row type, or collection data type.

element type

The data type of the elements in a collection.

embedded SQL

The SQL statements that are placed within some other host language. For example, Informix supports embedded SQL in C.

enabled mode

The default object mode of database objects. When a constraint, index, or trigger is in this mode, the database server recognizes the existence of the object and takes the object into consideration while executing data manipulation statements. See also object mode.

end-user format

The format in which data appears within a client application as literal strings or character variables. End-user formats are useful for data types whose database format is different from the format to which users are accustomed.

end-user routine

A user-defined routine (UDR) that performs a task within an SQL statement that the existing built-in routines do not perform. Examples of tasks include encapsulating multiple SQL statements, creating trigger actions, and restricting who can access database objects.

environment variable

A variable that the operating system maintains for each user and makes available to all programs that the user runs.

error log

A file that receives error information whenever a program runs.

error message

A message that is associated with a (usually negative) number. IBM Informix applications display error messages on the screen or write them to files.

error trapping

See exception handling.

escape character

A character that indicates that the next character, normally interpreted by the program as having special significance, is to be processed as a literal character instead. The escape character precedes the special character (such as a wildcard or delimiter) to "escape" (that is, ignore) the special significance.

escape key

The physical key of a keyboard, usually marked ESC, that is used to terminate one mode and start another mode in most UNIX and DOS systems.

ESQL/C Smart Large-Object API

An API of C routines that an IBM Informix ESQL/C client application can use to access smart large objects as operating-system files. The ESQL/C Smart Large-Object API is part of the IBM Informix ESQL/C SQL API. You can also access smart large objects with a set of functions in the DataBlade API.

exception

An error or warning that the database server returns, or a state that a SPL statement initiates.

exception handling

The code in a program that anticipates and reacts to runtime errors and warnings. Also referred to as error handling or error trapping.

exclusive access

Sole access to a database or table by a user. Other users are prevented from using it.

exclusive lock

A lock on an object (row, page, table, or database) that is held by a single thread that prevents other processes from acquiring a lock of any kind on the same object.

executable file

A file that contains code that can be executed directly. A C-language object file can be an executable file; it contains the machine-level instructions that correspond to the C-language source file.

execute

To run a statement, program, routine, or a set of instructions. See also executable file.

explicit cast

A user-defined cast that a user explicitly invokes with the CAST AS keyword or cast operator ( :: ). See also implicit cast.

explicit connection

A connection made to a database environment that uses the CONNECT statement. See also implicit connection.

explicit projection list

In a SELECT statement, a projection list that does not use the asterisk ( * ) notation, but explicitly lists the names of the columns from which the query returns values. (The projection list is sometimes called the select list.)

explicit transaction

A transaction that is initiated by the BEGIN WORK statement. This type of transaction is available only in non-ANSI compliant databases that support logging. See also implicit transaction and singleton implicit transaction.

exponent

The power to which a value is to be raised.

express mode

An Extended Parallel Server method of loading or unloading data that uses light appends.

expression

A specification that the database server can evaluate. This can include literal values, constants, column values, functions, quoted strings, operators, and parentheses (as delimiters). A Boolean expression evaluates as TRUE, FALSE, or UNKNOWN. An arithmetic expression can contain the arithmetic operators (+, -, ¥, /, and so on) and returns a number.

expression-based fragmentation

A distribution scheme that distributes rows to fragments according to a user-specified expression that is defined in the WHERE clause of an SQL statement.

extended data type

A term used to refer to data types that are not built in; namely complex data types, opaque data types, and distinct data types.

extent

A continuous segment of disk space that a database server allocated to a tblspace (a table). The user can specify both the initial extent size for a table and the size of all subsequent extents that a database server allocates to the table.

external function

An external routine that returns a single value.

external procedure

An external routine that does not return a value.

external routine

A user-defined routine that is written in an external language that the database supports. These external languages include the C and Java languages. The routine names, parameters, and other information are registered in the system catalog tables of a database. However, the executable code of an external routine is stored outside the database. An external routine can be an external function or an external procedure.

external space

Storage space that a user-defined access method manages rather than the database server. The IN clause of the CREATE TABLE and CREATE INDEX statements can specify the name of an external space instead of a dbspace.

external table

A database table that is not in the current database. It might or might not be in a database that the same database server manages.

extspace

(Not for Extended Parallel Server) A logical name associated with an arbitrary string that signifies the location of external data. Access its contents with a user-defined access method.

family name

A quoted string constant that specifies a family name in the optical family. See also optical family.

fault tolerance

See high availability.

fetch

The action of moving a cursor to a new row and retrieving the row values into memory.

fetch buffer

A buffer in the application process that the database server uses to send fetched row data (except TEXT and BYTE data) to the application.

field

A component of a named row type or unnamed row type that contains a name and a data type and can be accessed in an SQL statement by using dot notation in the form row type name.field name. See also column.

file

A collection of related information stored together on a system, such as the words in a letter or report, a computer program, or a listing of data.

file server

A network node that manages a set of disks and provides storage services to computers on the network.

filename extension

The part of a filename following the period. For example, DB–Access appends the extension .sql to command files.

filter

A set of conditions (sometimes called a predicate) for selecting rows or records. In an SQL query, the conditional expression in the WHERE clause is a filter that controls the active set of the query. The High-Performance Loader (HPL) uses a filter component to screen data before loading it into a database.

filtering mode

An object mode of a database object, causing bad rows to be written to the violations table during DML operations. During DML operations, the database server enforces requirements of a constraint or of a unique index that is in this mode and identifies any rows that would violate the requirement.

fixchar

A character data type in ESQL/C programs, for fixed-length character strings that are padded (as needed) with trailing blanks, and not NULL-terminated.

fixed-point number

A number where the decimal point is fixed at a specific place regardless of the value of the number.

flag

A command-line option, usually indicated by a minus ( - ) sign in UNIX systems. For example, in DB–Access the -e flag echoes input to the screen.

flexible temporary table

An explicit temporary table that Extended Parallel Server automatically fragments using a round-robin distribution scheme.

floating-point number

A number with fixed precision (total number of digits) and undefined scale (number of digits to the right of the decimal point). The decimal point floats as appropriate to represent an assigned value.

foreign key

A column or set of columns that references a unique or primary key in a table. For every entry in a foreign-key column containing only non-NULL values, there must exist a matching entry in the unique or primary column.

format

A description of the organization of a data file. See also component.

formatting character

For XPS, a percent sign (%) followed by a letter (c, n, o, or r). In a command line, Extended Parallel Server expands the formatting character to designate multiple coserver numbers (%c), multiple nodes (%n), multiple ordinal numbers designating dbspaces (%d), or a range of dbspaces (%r).

fragment

See index fragment and table fragment.

fragment elimination

The process of applying a filter predicate to the fragmentation strategy of a table or index and removing the fragments that do not apply to the operation.

fragmentation

The process of defining groups of rows within a table based on a rule and then storing these groups, or fragments, in separate dbspaces that you specify when you create a table or index fragmentation strategy.

full checkpoint

A type of checkpoint where the pages on disk are synchronized with the pages in the shared-memory buffer pool.

function

A routine that returns one or more values. See also user-defined function.

function cursor

A cursor that is associated with an EXECUTE FUNCTION statement, which executes routines that return values. See also cursor function.

function overloading

See routine overloading.

fuzzy checkpoint

A type of checkpoint where only certain pages on disk are synchronized with the pages in the shared-memory buffer pool, and the logical log is used to synchronize the rest of the pages during fast recovery.

gateway

A device that establishes data communications between networks.

generalized-key (GK) index

A type of index for static tables with Extended Parallel Server that can speed certain queries by storing the result of an expression as a key in a B+ tree or bitmap index. Three types of GK index are selective, virtual column, and join.

gigabyte

A unit of storage, equal to 1024 megabytes or 1024³ bytes.

Global Language Support (GLS)

A feature that enables Informix APIs and database servers to support different languages, cultural conventions, and code sets. For information about the GLS feature, see the IBM Informix: GLS User's Guide.

global variable

A variable or identifier whose scope of reference is all modules of a program. Compare with local variable.

globally detached index

For Extended Parallel Server, a type of index that has a fragmentation strategy that is independent of the table fragmentation and where the database server cannot verify that each index row resides on the same coserver as the referenced data row. You can use an expression, system-defined hash, or hybrid distribution scheme to create globally detached indexes for any table. See also locally detached index.

GLS

See Global Language Support (GLS).

GLS API

A legacy acronym for IBM Informix GLS. An API of C routines that a C-language external routine can use to access IBM Informix GLS locales. This API also includes functions that obtain culture-specific collation, and time, date, number, and currency formats and functions that provide a uniform way of accessing character data, regardless of whether the locale is left-to-right or bidirectional or supports single-byte characters or multibyte characters.

hash fragmentation

See system-defined hash fragmentation.

hash rule

A user-defined algorithm that maps each row in a table to a set of hash values and that is used to determine the fragment in which a row is stored.

header file

A source file that contains declarations for variables, constants, and macros that a particular group of modules or programs share.

help message

Online text displayed automatically or at the request of the user to assist the user in interactive programs. Such messages are stored in help files.

heterogeneous commit

A protocol governing a group of database servers, of which at least one is a gateway participant. It ensures the all-or-nothing basis of distributed transactions in a heterogeneous environment. See also two-phase commit.

hierarchy

A tree-like data structure in which some groups of data are subordinate to others such that only one group (called root) exists at the highest level, and each group except root is related to only one parent group on a higher level.

high availability

The ability of a system to resist failure and data loss. High availability includes features such as fast recovery and mirroring. It is sometimes referred to as fault tolerance.

High-Performance Loader

The High-Performance Loader (HPL) utility is part of Dynamic Server. The HPL loads and unloads data using parallel access to devices. See also external table.

highlight

A rectangular inverse-video area marking your place on the screen. A highlight can indicate the current option on a menu or the current character in an editing session. If a terminal cannot display highlighting, the current option can appear in angle ( < > ) brackets, with the current character underlined.

hold cursor

A cursor that is created using the WITH HOLD keywords. A hold cursor remains open past the end of a transaction. It allows uninterrupted access to a set of rows across multiple transactions.

home page

The page that contains the first byte of the data row, specified by the rowid. Even if a data row outgrows its original storage location, the home page does not change. The home page contains a forward pointer to the new location of the data row. See also remainder page.

host variable

An SQL API program variable that you use in an embedded statement to transfer information between the SQL API program and the database.

HPL

See High-Performance Loader.

hybrid fragmentation

A distribution scheme that lets the user specify two fragmentation methods. Usually one method is used globally and one method is used locally.

identifier

In the default locale, a sequence of letters, digits, and underscores (_) that is the unqualified name of a database, storage, or program object. (Additional characters are valid in other locales or if the DELIMIDENT variable is set.)

implicit transaction

A transaction that begins implicitly after the COMMIT WORK or ROLLBACK WORK statement. This is the only type of transaction that ANSI-compliant databases support, but it is also available for other databases that support logging. See also explicit transaction and singleton implicit transaction.

implicit cast

A built-in or user-defined cast that the database server automatically invokes to perform data-type conversion. See also explicit cast.

implicit connection

A connection that is made using the DATABASE, CREATE DATABASE, START DATABASE, or DROP DATABASE statement. See also explicit connection.

implicit projection list

In a SELECT statement, using the asterisk ( * ) wildcard symbol so that a query returns values from all columns of the table. The projection list is sometimes called the "select list," because it immediately follows the SELECT keyword.

incremental archiving

A system of archiving that allows the option to archive only those parts of the data that have changed since the last archive was created.

independent subquery

A subquery that has no relationship to or dependency on any of its parent queries. It needs to be evaluated only once and the results can be used thereafter. In independent subqueries, both the parent and subquery are parallelized. See also correlated subquery and subquery.

index

A structure of entries, each of which contains a value or values and a pointer to the corresponding location in a table or smart large object. An index might improve the performance of database queries by ordering a table according to key column values or by providing access to data inside of large objects.

index fragment

Consists of zero or more index items grouped together, which can be stored in the same dbspace as the associated table fragment or in a separate dbspace. An index fragment also might occupy an sbspace or an extspace.

Informix user ID

A login user ID (login user name or authorization identifier) that must be valid on all computer systems (operating systems) involved in the client's database access. Often referred to as the client's user ID. This need not refer to a fully functional user account on the computer system; only the user identity components of the user account information are significant to Informix database servers. A given user typically has the same Informix user ID on all networked computer systems involved in the database access. The authorization identifier informix is required for some database objects and operations.

Informix user password

A user ID password that must be valid on all computer systems (operating systems) involved in the client's database access. When the client specifies an explicit user ID, most computer systems require the Informix user password.

inheritance

The process that allows an object to acquire the properties of another object. Inheritance allows for incremental modification, so that an object can inherit a general set of properties and add properties that are specific to itself. See also type inheritance and table inheritance.

initialize

To prepare for execution. To initialize a variable, you assign it a starting value. To initialize the database server, you start its operation.

inmigration

The process by which Optical Subsystem migrates TEXT and BYTE data from the optical storage subsystem into the Dynamic Server environment.

inner join

A query that symmetrically combines rows from two or more tables. See also simple inner join.

input

The information that is received from an external source (for example, from the keyboard, a file, or another program) and processed by a program.

input parameter

In a prepared SQL statement, a value, represented by a "?" placeholder symbol, that must be provided when the prepared statement is executed.

insert cursor

A cursor for insert operations, associated with an INSERT statement. Allows bulk insert data to be buffered in memory and written to disk.

installation

The loading of software from some medium (tape, cartridge, removable disk, or CD) onto a computer, and preparing the software for use.

interactive

An aspect of a program that accepts input from the user, processes the input, and then produces output on the screen, in a file, or on a printer.

internationalization

The process of making IBM Informix products easily adaptable to the conventions of various cultures and natural languages. Among other features, internationalized software can support culturally-specific collation and date, time, and money formats. See also IBM Informix Guide to GLS Functionality.

interquery parallelization

The ability to process multiple queries simultaneously to avoid a performance delay when multiple independent queries access the same table. See also intraquery parallelization.

interrupt

A signal from a user or another process that can stop the current process temporarily or permanently. See also signal.

interrupt key

A key used to cancel or abort a program or to leave a current menu and return to the menu one level above. On many systems, the interrupt key is CONTROL-C; on some others, the interrupt key is DEL or CONTROL-Break.

intraquery parallelization

Breaking of a single query into subqueries by a database server using PDQ and then processing the subqueries in parallel. Parallel processing of this type has important implications when each subquery retrieves data from a fragment of a table. Because each partial query operates on a smaller amount of data, the retrieval time is significantly reduced and performance is improved. See also interquery parallelization.

IPX/SPX

Acronym for Internetwork Packet Exchange/Sequenced Packet Exchange. It refers to the NetWare network protocol by Novell.

ISAM

Acronym for Indexed Sequential Access Method. This allows you to find information in a specific order or to find specific items of information quickly through an index. See also access method.

ISAM error

Operating system or file access error.

ISO

Acronym for the International Standards Organization. ISO sets worldwide standards for the computer industry, including standards for character input and manipulation, code sets, and SQL syntax.

ISO8859-1

A code set that contains 256 single-byte characters. Characters 0 through 127 are the ASCII characters. Characters 128 through 255 are mostly non-English characters from European languages; for example, é, ñ and ö.

isolation

When multiple users attempt to access common data, the level of independence specifically relating to the locking strategy for read-only SQL requests. The various levels of isolation are distinguished primarily by the length of time that shared locks are (or can be) acquired and held. You can set the isolation level with the SET ISOLATION or SET TRANSACTION statement.

iterator function

A cursor function that returns a data set by successive calls. It can be written in C or Java or can be an SPL routine that includes RETURN WITH RESUME.

jagged rows

A query result in which rows differ in the number and type of columns they contain because the query applies to more than one table in a table hierarchy.

join

The process of combining information from two or more tables based on some common domain of information. Rows from one table are paired with rows from another table when information in the corresponding rows match on the joining criterion. For example, if a customer_num column exists in the customer and the orders tables, you can construct a query that pairs each customer row with all the associated orders rows based on the common customer_num. See also Cartesian product and outer join.

join index

A type of generalized-key (GK) index that contains keys that are the result of a query that joins multiple tables.

jukebox

A physical storage device that consists of one or more optical-disc drives, slots for platters, optical platters, and a robotic arm to transfer platters between the slots and the drives. A jukebox is also known as an autochanger.

kernel

Operating system component that controls processes and resource allocation.

key

The items of information that are used to locate a row of data. A key defines the pieces of information for which you want to search as well as the order in which you want to process information in a table. For example, you can index the last_name column in a customer table to find specific customers or to process the customers in alphabetical order (or in reverse alphabetical order) by their last names (when last_name serves as the key).

keyword

A word that has a predefined meaning in a programming language. For example, the word SELECT is a keyword in SQL.

kilobyte

A unit of storage that consists of 1024 bytes.

Language Supplement

An IBM Informix product that provides the locale files and error messages to support one or more languages. The International Language Supplement supports several European languages. IBM Informix provides separate Language Supplements for several Asian languages.

large object

A data object that is logically stored in a column of a table, but physically stored independently of the column, due to its size. Large objects can be simple large objects (TEXT, BYTE) or smart large objects (BLOB, CLOB).

leaf node

Index page containing index items and horizontal pointers to other leaf nodes. A database server creates leaf nodes when the root node becomes full.

level of isolation

See isolation.

library

A group of precompiled routines designed to perform tasks that are common to a given kind of application. An application programming interface (API) can include a library of routines that you can call from application programs. See also dynamic link library (DLL), shared library, and shared-object file.

light append

An unbuffered, unlogged insert operation.

link

A way of combining separately compiled program modules, usually into an executable program. Compare with compile and execute.

LIST

A collection data type created with the LIST constructor in which elements are ordered and duplicates are allowed.

literal

The representation of a data type value in a format that the database server accepts in data-entry operations. For example, 234 is a literal integer and "abcd" is a literal character.

little-endian

A hardware-determined storage method in which the least-significant byte of a multibyte number has the lowest address. See also big-endian.

load job

The information required to load data into a relational database using the HPL. This information includes the format, map, filter, device array, project, and special options.

local copy

For Extended Parallel Server, a replica of a table on a local coserver that is copied to multiple coservers. This allows faster access to the data for OLTP transactions connected to those coservers because you do not have to send the data across the communication links between coservers.

local loopback

A connection between the client application and database server that uses a network connection even though the client application and the database server are on the same computer.

local variable

A variable or identifier whose scope of reference is only within the routine in which it is defined. Compare with global variable.

locale

A set of Informix files that specify the linguistic conventions for a country, region, culture, or language. IBM Informix products provide predefined locales that customers cannot modify. A locale provides the name of the code set that the application data uses, the collation order to use for character data, and the end-user format. See also client locale, database locale, default locale, server locale, and server-processing locale.

localized order

A collation order other than code-set order, if defined for a locale. Only NCHAR and NVARCHAR data values are collated in a localized order. Database objects collate in their creation-time order, if this is not the runtime order.

locally-detached index

For Extended Parallel Server, a type of index that has a fragmentation strategy that is independent of the table fragmentation but where the database server recognizes that each index row resides on the same co-server as the referenced data row. You can use an expression, system-defined hash, or hybrid distribution scheme to create locally detached indexes for any table. See also globally-detached index.

lock coupling

A locking feature that holds a lock on the child node until a lock is obtained on the parent node during upward movement when updating an R-tree index. Lock coupling is used if the bounding box of a leaf node has changed. You must propagate the change to the parent node by moving upwards in the tree until you reach a parent node that needs no change.

lock mode

An option to specify whether a user who requests a lock on an already locked object waits until the object is released to receive the lock, or immediately receives an error, or waits for some span of time before receiving an error.

locking

A concurrency feature temporarily limiting access to an object (database, table, page, or row) to prevent conflicting interactions among concurrent processes. An exclusive lock restricts read and write access to only one user; a shared lock allows read-only access to other users. An update lock begins as a shared lock, but is upgraded to an exclusive lock after a row is changed.

locking granularity

The size of a locked object. The size can be a database, table, page, or row.

logical log

An allocation of disk space that the database server manages that contains records of all changes that were performed on a database during the period when the log was active. It is used to roll back transactions, recover from system failures, and restore databases from archives. See also physical log.

login

The process of identifying oneself to a computer.

login password

See Informix user password.

login user ID

See Informix user ID.

logslice

A dbslice whose contents are comprised solely of logical-log files. The logical-log files in the logslice can be owned by multiple coservers, one log file per dbspace. See also dbslice, rootslice, and physslice.

LVARCHAR

A built-in data type that stores varying-length character data of up to 32 kilobytes.

macro

A named set of instructions that the computer substitutes when it encounters the name in source code.

mantissa

The significant digits in a floating-point number.

map

A description of the relation between the records of a data file and the columns of a relational database. See also component.

massively parallel processing system

A system composed of multiple computers that are connected to a single high-speed communication subsystem. MPP computers can be partitioned into nodes. Compare with symmetric multiprocessing system.

megabyte

A unit of storage that equals 1024 kilobytes or 1024² bytes.

Memory Grant Manager (MGM)

(Not for Extended Parallel Server) A database server component that coordinates the use of memory and I/O bandwidth for decision-support queries. MGM uses the DS_MAX_QUERIES, DS_TOTAL_MEMORY, DS_MAX_SCANS, and PDQPRIORITY configuration parameters to determine what resources can or cannot be granted to a decision-support query.

menu

A screen display that allows you to choose the commands that you want the computer to perform.

MGM

Acronym for Memory Grant Manager.

mirroring

Storing the same data on two chunks simultaneously. If one chunk fails, the data values are still usable on the other chunk in the mirrored pair. The database server administrator handles this data storage option.

MODE ANSI

The keywords specified on the CREATE DATABASE or START DATABASE statement to make a database ANSI compliant.

monochrome

A term that describes a monitor that can display only one color.

MPP

Acronym for massively parallel processing system.

multibyte character

A character that might require from two to four bytes of storage. If a language contains more than 256 characters, the code set must contain multibyte characters. With a multibyte code set, an application cannot assume that one character requires only one byte of storage. See also single-byte character.

multiplexed connection

A single network connection between a database server and a client application that handled multiple database connections from the client.

MULTISET

A collection data type created with the MULTISET constructor in which elements are not ordered and duplicates are allowed.

multithreading

Running of multiple threads within the same process. See thread.

named row type

A row type created with the CREATE ROW TYPE statement that has a declared name and inheritance properties and can be used to construct a typed table.

national character

In National Language Support (NLS), a character from the written form of a natural language that can be stored in an NCHAR or NVARCHAR column. Sometimes called a native character.

native character

See national character.

National Language Support (NLS)

A feature that supports single-byte code sets, using NCHAR and NVARCHAR columns to store non-English character strings. This technology has been superseded on IBM Informix databases by Global Language Support (GLS).

NLS

See National Language Support (NLS).

node

(1) In the context of an index for a database, a node is an ordered group of key values having a fixed number of elements. (A key is a value from a data record.) A B-tree for example, is a set of nodes that contain keys and pointers arranged in a hierarchy. See also branch node, leaf node, and root node.

(2) In the context of a MPP system, a node is an individual computer. See also massively parallel processing system.

(3) In the context of a SMP system, a node can either be the entire SMP computer or a fully functioning subsystem that uses a portion of the hardware resources of that SMP system. See also symmetric multiprocessing system.

(4) For Extended Parallel Server, a node is an individual computer with one or more CPUs that runs a single instance of an operating system within a parallel-processing platform. A node can be a uniprocessor, a cluster of stand-alone computers, an SMP computer, or an independent subsystem configured within an SMP computer.

non-ASCII character

A character that is not match any of the 128 code points in the ASCII character set. Non-ASCII characters include 8-bit characters and multibyte characters.

noncursor function

A user-defined function that returns a single group of values (one row of data) and therefore does not require a cursor when it is executed. Compare with cursor function.

nonvariant function

A user-defined function that always returns the same value when passed the same arguments. A nonvariant function must not contain SQL statements. Compare with variant function.

NOT NULL constraint

A constraint on a column that specifies that the column cannot contain NULL values.

NULL

A keyword of SQL, indicating that a value is unknown, missing, or not applicable. (A NULL value is not the same as a value of zero or blank.)

object

See database object.

object mode

The state of a database object as recorded in the sysobjstate system catalog table. A constraint or unique index can be in enabled, disabled, or filtering mode. A trigger or duplicate index can be in enabled or disabled mode. You use SET statements to change the mode of an object.

object-relational database

A database that adds object-oriented features to a relational database, including support for user-defined data types, user-defined routines, user-defined casts, user-defined access methods, and inheritance.

OLTP

Acronym for Online Transaction Processing. See also online transaction processing application.

online transaction processing application

Characterized by quick, indexed access to a small number of data items. The applications are typically multiuser, and response times are measured in fractions of seconds. See also decision-support application.

online transaction processing queries

The transactions that OLTP applications handle are usually simple and predefined. A typical OLTP system is an order-entry system where only a limited number of rows are accessed by a single transaction many times. See also decision-support query.

opaque data type

A data type whose inner structure is not visible to the database server. Opaque types that are not built-in need user-defined routines and user-defined operators that work on them. Synonym for base type and user-defined base type.

opaque-type support function

One of a group of user-defined functions that the database server uses to perform operations on opaque data types (such as converting between the internal and external representations of the type).

open

The process of making a resource available, such as preparing a file for access, activating a column, or initiating a window.

operational table

A logging permanent table that uses light appends for fast update operations. Operational tables do not perform record-by-record logging.

operator

In an SQL statement, a keyword (such as UNITS or UNION) or a symbol (such as =, >, <,+, -, or *) that invokes an operator function. The operands to the operator are arguments to the operator function.

operator binding

The implicit invocation of an operator function when an operator is used in an SQL statement.

operator class

An association of operator-class functions with a secondary access method. The database server uses an operator class to optimize queries and build an index of that secondary access method.

operator-class function

One of the operator-class support functions or operator-class strategy functions that constitute an operator class. For user-defined operator classes, the operator-class functions are user-defined functions.

operator-class strategy function

An operator-class function that can appear as a filter in a query. The query optimizer uses the strategy functions to determine if an index of a particular secondary access method can be used to process the filter. You register operator-class strategy functions in the STRATEGIES clause of the CREATE OPCLASS statement.

operator-class support function

An operator-class function that a secondary access method uses to build or search an index. You register operator-class support functions in the SUPPORT clause of the CREATE OPCLASS statement.

operator function

A function that processes one or more arguments (its operands) and returns a value. Many operator functions have corresponding operators, such as plus() and +. You can overload an operator function so that it handles a user-defined data type. See also routine overloading.

optical cluster

An amount of space on an optical disc that is reserved for storing a group of logically related simple large objects or smart large objects.

optical family

A group of optical discs, theoretically unlimited in number.

optical platter

A removable optical disc that stores data in an optical storage subsystem.

optical statements

The SQL statements that you use to control optical clustering.

optical volume

One side of a removable Write-Once-Read-Many (WORM) optical disc.

outer join

An asymmetric joining of a dominant (outer) table and one or more "subservient" tables in a query, where the values for the outer table are preserved even if no matching rows exist in the subservient table. In the result set, any outer-table row that has no matching row in the subservient table has NULL values in the columns selected from the subservient table.

outmigration

The process by which Optical Subsystem migrates TEXT or BYTE data from the Dynamic Server environment to an optical storage subsystem.

output

The result that the computer produces in response to a query or a request for a report.

overloading

See routine overloading.

owner-privileged

A class of SPL routines that any user can create who has Resource database privileges.

packed decimal

A storage format that represents either two decimal digits or a sign and one decimal digit in each byte.

pad

Usually, to fill empty places at the beginning or end of a line, string, or field with spaces or blanks when the input is shorter than the field.

page

The physical unit of disk storage and basic unit of memory storage that the database server uses to read from and write to Informix databases. Page size is fixed for a particular operating system and platform. A page is always entirely contained within a chunk. See also home page and remainder page.

parallel database query

The execution of SQL queries in parallel rather than sequential order. The tasks a query requires are distributed across several processors. This type of distribution enhances database performance.

parallel-processing platform

A parallel-processing platform is a set of independent computers that operate in parallel and communicate over a high-speed network, bus, or interconnect. See also symmetric multiprocessing system and massively parallel processing system.

parallelism

Ability of an Informix database server to process a task in parallel by breaking the task into subtasks and processing the subtasks simultaneously, thus improving performance.

parameter

A variable that is given a value for a specified application. In the signature of a user-defined routine, a parameter serves as a placeholder for an argument. The parameter specifies the data type of the value that the user-defined routine expects when it receives the associated argument at runtime. See also configuration file, input parameter, and routine signature.

parent-child relationship

See referential constraint.

parent table

The referenced table in a referential constraint. See also child table.

participating coserver

A coserver that controls one or more fragments of a table that Extended Parallel Server accesses. See also coserver and connection coserver.

partition

See table fragment.

pattern

An identifiable or repeatable series of characters or symbols.

PDQ

Acronym for parallel database query.

PDQ priority

Determines the amount of resources that a database server allocates to process a query in parallel. These resources include memory, threads (such as scan threads), and sort space. The level of parallelism is established by using the PDQPRIORITY environment variable or various database server configuration parameters (including PDQPRIORITY and MAX_PDQPRIORITY) or dynamically through the SET PDQPRIORITY statement.

permission

On some operating systems, the right to access files and directories.

phantom row

A row of a table that is initially modified or inserted during a transaction but is subsequently rolled back. Another user can see a phantom row if the isolation level is Informix Dirty Read or ANSI compliant Read Uncommitted. No other isolation level allows a user to see a changed but uncommitted row.

physical log

A set of contiguous disk pages in shared memory where the database server stores an unmodified copy (before-image) of pages before the changed pages are recorded. The pages in the physical log can be any database server page except a blobspace blobpage.

physslice

A dbslice that contains the physical log. See also dbslice, logslice, and rootslice.

pointer

A value that specifies the address in memory of the data or variable, rather than the contents of the data or variable.

polymorphism

See routine overloading and type substitutability.

precision

The total number of significant digits in a real number, both to the right and left of the decimal point. For example, the number 1437.2305 has a precision of 8. See also scale.

predefined opaque data type

An opaque data type for which the database server provides the data type definition. See also BLOB, BOOLEAN, CLOB, LVARCHAR, and pointer.

predicate

See filter.

predicate lock

A lock held on index keys that qualifies for a predicate. In a predicate lock, exclusive predicates consist of a single key value, and shared predicates consist of a query rectangle and a scan operation such as inclusion or overlap.

prepared statement

An SQL statement that is generated by the PREPARE statement from a character string or from a variable that contains a character string. This feature allows you to form your request while the program is executing without having to modify and recompile the program.

preprocessor

A program that takes high-level programs and produces code that a standard language compiler such as C can compile.

primary access method

An access method whose routines access a table with such operations as inserting, deleting, updating, and scanning. See also secondary access method.

primary key

The information from a column or set of columns that uniquely identifies each row in a table. The primary key sometimes is called a unique key.

primary-key constraint

Specifies that each entry in a column or set of columns contains a unique non-NULL value.

printable character

A character that can be displayed on a terminal, screen, or printer. Printable characters include A-Z, a-z, 0-9, and punctuation marks. Compare with control character.

privilege

The right to use or change the contents of a database, table, table fragment, or column. See also access privileges.

procedure

A routine that does not return values. See also user-defined procedure.

procedure overloading

See routine overloading.

process

A discrete task, generally a program, that the operating system executes.

project

A group of related components that the High-Performance Loader (HPL) uses. See also component.

projection

Taking a subset of the columns in a table. Projection is implemented through the Projection clause of the SELECT statement and returns some of the columns and all the qualifying rows of a table. The set of columns in the Projection clause is sometimes called the "select list." See also selection and join.

promotable lock

A lock that can be changed from a shared lock to an exclusive lock. See also update lock.

protocol

A set of conventions that governs communication among computers. These conventions govern format, timing, sequencing, and error control.

query

A request to the database to retrieve data that meet certain criteria, usually made with the SELECT statement.

query optimization information statements

The SQL statements that are used to optimize the performance of queries. These statements include SET EXPLAIN, SET OPTIMIZATION, and UPDATE STATISTICS.

query unnesting

An execution strategy for nested SQL subqueries whereby Extended Parallel Server rewrites such subqueries to use modified joins rather than iteration mechanisms. The sqexplain.out file reflects the query plan that has been selected after subquery unnesting has occurred.

R-tree index

(Not for Extended Parallel Server) A type of index that uses a tree structure based on overlapping bounding rectangles to speed access to spatial and multidimensional data types. See also bitmap index and B-tree index.

range fragmentation

A distribution scheme that distributes data in table fragments that contain a specified key range. This technique can eliminate scans of table fragments that do not contain the required rows, making queries faster.

range rule

A user-defined algorithm for expression-based fragmentation. It defines the boundaries of each fragment in a table using SQL relational and logical operators. Expressions in a range rule can use the following restricted set of operators: >, <, >=, <=, and the logical operator AND.

raw device

See unbuffered disk I/O.

raw disk

See unbuffered disk I/O.

raw table

A nonlogged permanent table that uses light appends.

Read Committed

A level of isolation that the SET TRANSACTION statement can specify, in which a user can view rows that are currently committed at the moment of the query request, but cannot view rows that were changed as part of a currently uncommitted transaction. This is the default isolation level for databases that are not ANSI compliant. See also Committed Read.

Read Uncommitted

An ANSI-compliant level of isolation, set with the SET TRANSACTION statement, that does not account for locks. This allows a user to view any existing rows, even rows that can be altered within currently uncommitted transactions. Read Uncommitted is the lowest level of isolation (no isolation at all), and is thus the most efficient. See also Dirty Read.

real user ID

See Informix user ID.

record

See row.

Record-ID

A four-byte RSAM entity, also known as RID, that describes the logical position of the record within a fragment. Not the same as rowid.

recover a database

To restore a database to a former condition after a system failure or other destructive event. The recovery restores the database as it existed immediately before the failure.

referential constraint

The relationship between columns within a table or between tables; also known as a parent-child relationship. Referencing columns are also known as foreign keys.

registering

In a database, the process of storing information about a database object in the system catalog tables of a database. Most SQL data definition statements perform some type of registration. For example, the CREATE FUNCTION and CREATE PROCEDURE statements register a user-defined routine in a database.

relation

See table.

relational database

A database that uses table structures to store data. Data in a relational database is divided across tables in such a way that additions and modifications to the data can be made easily without loss of information.

relational database server

A database server that manages data values that are stored in rows and columns.

remainder page

A page that accommodates subsequent bytes of a long data row. If the trailing portion of a data row is less than a full page, it is stored on a remainder page. After the database server creates a remainder page for a long row, it can use the remaining space in the page to store other rows. Each full page that follows the home page is referred to as a big-remainder page.

remote connection

A connection that requires a network.

remote routine

A routine in a databases of a remote server. See subordinate server.

remote server

See subordinate server.

remote table

In a distributed query, a table in a database of a server that is not the local database server. See also coordinating server, subordinate server.

Repeatable Read

An Informix and ANSI level of isolation available with the Informix SET ISOLATION statement or the ANSI compliant SET TRANSACTION statement, which ensures that all data values read during a transaction are not modified during the entire transaction. Transactions under ANSI Repeatable Read are also known as Serializable. Informix Repeatable Read is the default level of isolation for ANSI compliant databases. See also isolation and Serializable.

reserved pages

The first 12 pages of the initial chunk of the root dbspace. Each reserved page stores specific control and tracking information that the database server uses.

reserved word

A word in a statement or command that you cannot use in any other context of the language or program without receiving a warning or error message.

restore a database

See recover a database.

role

A classification or work task, such as payroll, that the DBA assigns. Assignment of roles makes management of privileges convenient.

role separation

(Not for Extended Parallel Server) A database server installation option that allows different users to perform different administrative tasks.

roll back

The process that reverses an action or series of actions on a database. The database is returned to the condition that existed before the actions were executed. See also transaction and commit work.

root dbspace

The initial dbspace that the database server creates. It contains reserved pages and internal tables that describe and track all other dbspaces, blobspaces, sbspaces, tblspaces, chunks, and databases.

root node

A single index page that contains node pointers to branch nodes. The database server allocates the root node when you create an index for an empty table.

root supertype

The named row type at the top of a type hierarchy. A root supertype has no supertype above it.

rootslice

A dbslice that contains the root dbspaces for all coservers for Extended Parallel Server. See also dbslice, logslice, and physslice.

round-robin fragmentation

A distribution scheme in which the database server distributes rows sequentially and evenly across specified dbspaces.

routine

A group of program statements that perform a particular task. A routine can be a function or a procedure. All routines can accept arguments. See also built-in and user-defined routine.

routine modifier

A keyword in the WITH clause of a CREATE FUNCTION, CREATE PROCEDURE, ALTER FUNCTION, ALTER PROCEDURE, or ALTER ROUTINE statement that specifies a particular attribute or usage of a user-defined routine.

routine overloading

The ability to assign one name to multiple user-defined routines and specify parameters of different data types on which each routine can operate. An overloaded routine is uniquely defined by its routine signature.

routine resolution

The process that the database server uses to determine which user-defined routine to execute, based on the routine signature. See also routine overloading.

routine signature

The information that the database server uses to uniquely identify a user-defined routine. The signature includes the type of routine (function or procedure); the routine name; and the number, order, and data types of the parameters. See also routine overloading and specific name.

row

A group of related items of information about a single entity across all columns in a database table. In a table of customer information, for example, a row contains information about a single customer. A row is sometimes referred to as a record or tuple. In an object-relational model, each row of a table stands for one instance of the subject of the table, which is one particular example of that entity. In a screen form, a row can refer to a line of the screen.

row type

A complex data type that contains one or more related data fields, of any data type, that form a template for a record. The data in a row type can be stored in a row or column. See also named row type and unnamed row type.

row variable

An IBM Informix ESQL/C host variable or SPL variable that holds an entire row type and provides access to the individual fields of the row.

rowid

In nonfragmented tables, rowid refers to an integer that defines the physical location of a row. Rowids must be explicitly created to be used in fragmented tables, and they do not define a physical location for a row. Rowids in fragmented tables are accessed by an index that is created when the rowid is created; this access method is slow. It is recommended that users creating new applications move toward using primary keys as a method of row identification instead of using rowids.

rule

How a database server or a user determines into which fragment rows are placed. The database server determines the rule for round-robin fragmentation and system-defined hash fragmentation. The user determines the rule for expression-based fragmentation and hybrid fragmentation. See also arbitrary rule and range rule.

runtime environment

The hardware and operating-system services available at the time a program runs.

runtime error

An error that occurs during program execution. Compare with compile-time error.

sbspace

(Not for Extended Parallel Server) A logical storage area that contains one or more chunks that store only BLOB and CLOB data.

scale

The number of digits to the right of the decimal place in DECIMAL notation. The number 14.2350 has a scale of 4 (four digits to the right of the decimal point). See also precision.

scale up

The ability to compensate for an increase in query size by adding a corresponding amount of computer resources so that processing time does not also increase.

scan thread

A database server thread that is assigned the task of reading rows from a table. When a query is executed in parallel, the database server allocates multiple scan threads to perform the query in parallel.

schema

The structure of a database or a table. The schema for a table lists the names of the columns, their data types, and (where applicable) the lengths, indexing, and other information about the structure of the table.

scope of reference

The portion of a routine or application program where an identifier can be accessed. Three possible scopes exist: local (applies only in a single statement block), modular (applies throughout a single module), and global (applies throughout the entire program). See also local variable and global variable.

scroll cursor

A cursor created with the SCROLL keyword that allows you to fetch rows of the active set in any sequence.

secondary access method

An access method whose routines access an index with such operations as inserting, deleting, updating, and scanning. See also operator class and primary access method.

secure auditing

A facility of Informix database servers that lets a database system security officer (DSSO) monitor unusual or potentially harmful user activity. Use the onaudit utility to enable auditing of events and to create audit masks. Use the onshowaudit utility to extract audit event data for analysis. (For more details of secure auditing, see the IBM Informix: Trusted Facility Guide.)

select

See query.

select cursor

A cursor that is associated with a SELECT statement, which lets you scan multiple rows of data, moving data row by row into a set of receiving variables.

selection

A horizontal subset of the rows of a single table that satisfies a specified condition. Selection is implemented through the WHERE clause of a SELECT statement and returns some of the rows and all of the columns in a table. See also projection and join.

selective index

A type of generalized-key index that contains keys for only a subset of a table.

selectivity

The proportion of rows within the table that a query filter can pass.

self-join

A join between a table and itself. A self-join occurs when a table is used two or more times in a SELECT statement (with different aliases) and joined to itself.

semaphore

An operating-system communication device that signals a process to awaken.

sequence

A database object (sometimes called a sequence generator) that can generate unique integer values in the INT8 range.

sequential cursor

A cursor that can fetch only the next row in sequence. A sequential cursor can read through a table only once each time the sequential cursor is opened.

Serializable

An ANSI-compliant level of isolation set with the SET TRANSACTION statement, ensuring all data read during a transaction is not modified during the entire transaction. See also isolation and Repeatable Read.

server locale

The locale that a database server uses when it performs its own read and write operations. The SERVER_LOCALE environment variable can specify a nondefault locale. See also client locale and locale.

server name

The unique name of a database server, assigned by the database server administrator, that an application uses to select a database server.

server number

A unique number between 0 and 255, inclusive, that a database server administrator assigns when a database server is initialized.

server-processing locale

The locale that a database server determines dynamically for a given connection between a client application and a database. See also locale.

session

The structure that is created for an application using the database server.

SET

A collection data type created with the SET type constructor, in which elements are not ordered and duplicate values can be inserted.

shared library

A shared-object file on a UNIX system. See also dynamic link library (DLL).

shared lock

A lock that more than one thread can acquire on the same object. Shared locks allow for greater concurrency with multiple users; if two users have shared locks on a row, a third user cannot change the contents of that row until both users (not just the first) release the lock. Shared-locking strategies are used in all levels of isolation except Informix Dirty Read and ANSI-compliant Read Uncommitted.

shared memory

A portion of main memory that is accessible to multiple processes. Shared memory allows multiple processes to communicate and access a common data space in memory. Common data does not have to be reread from disk for each process, reducing disk I/O and improving performance. Also used as an Inter-Process Communication (IPC) mechanism to communicate between two processes running on the same computer.

shared-object file

A library that is not linked to an application at compile time but instead is loaded into memory by the operating system as needed. Several applications can share access to the loaded shared-object file. See also dynamic link library (DLL) and shared library.

shelf

The location of an optical platter that is neither on an optical drive nor in a jukebox slot.

shuffling

Shuffling refers to the process that occurs when a database server moves rows or key values from one fragment to another. Shuffling occurs in a variety of circumstances including when you attach, detach, or drop a fragment.

signal

A means of asynchronous communication between two processes. For example, signals are sent when a user or a program attempts to interrupt or suspend the execution of a process.

signature

See routine signature.

simple inner join

A join that combines information from two or more tables based on the relationship between one column in each table. Rows that do not satisfy the join criteria are discarded from the result. See also composite join.

simple large object

A large object that is stored in a blobspace or dbspace is not recoverable and does not obey transaction isolation modes. Simple large objects include TEXT and BYTE data types. Extended Parallel Server does not support simple large objects that are stored in a blobspace.

simple predicate

A search condition in the WHERE clause that has one of the following forms: f(column, constant), f(constant, column), or f(column), where f is a binary or unary function that returns a Boolean value (TRUE, FALSE, or UNKNOWN).

single-byte character

A character that uses one byte of storage. Because a single byte can store values in the range of 0 to 255, it can uniquely identify 256 characters. With these code sets, an application can assume that one character is always stored in one byte. See also 8-bit character and multibyte character.

singleton implicit transaction

A single-statement transaction that does not require either a BEGIN WORK or a COMMIT WORK statement. This type of transaction can occur only in a database that is not ANSI compliant, but that supports transaction logging. See also explicit transaction and implicit transaction.

singleton select

A SELECT statement that returns a single row.

smart large object

A large object that is stored in an sbspace, which has read, write, and seek properties similar to a UNIX file, is recoverable, obeys transaction isolation modes, and can be retrieved in segments by an application. Smart large objects include BLOB and CLOB data types.

SMI

Acronym for system-monitoring interface.

SMP

See symmetric multiprocessing system.

source file

A text file that contains instructions in a high-level language, such as C. A C source file is compiled into an executable file called an object file. An SPL source file is compiled into its own executable format. See also compile.

source type

The data type from which a DISTINCT type is derived.

specific name

A name that you can assign to an overloaded user-defined routine to uniquely identify a particular signature of the user-defined routine. See also routine overloading and routine signature.

speed up

The ability to add computing hardware to achieve correspondingly faster performance for a DSS query or OLTP operation of a given volume.

SPL

See Stored Procedure Language (SPL).

SPL function

An SPL routine that returns one or more values.

SPL procedure

An SPL routine that does not return a value.

SPL routine

A user-defined routine that is written in Stored Procedure Language (SPL). Its name, parameters, executable format, and other information are stored in the system catalog tables of a database. An SPL routine can be an SPL procedure or an SPL function.

SPL variable

A variable that is declared with the DEFINE statement in an SPL routine.

SQL

Acronym for Structured Query Language. SQL is a database query language that was developed by IBM and standardized by ANSI. Informix relational database management products are based on an extended implementation of ANSI-standard SQL.

SQL API

An application programming interface that allows you to embed SQL statements directly in an application. The embedded-language product IBM Informix ESQL/C is an example of an SQL API. See also host variable.

SQLCA

Acronym for SQL Communications Area. The SQLCA is a data structure that stores information about the most recently executed SQL statement. The result code returned by the database server to the SQLCA is used for error handling by Informix SQL APIs.

sqlda

Acronym for SQL descriptor area. A dynamic SQL management structure that can be used with the DESCRIBE statement to store information about database columns or host variables used in dynamic SQL statements. The sqlda structure is an Informix-specific structure for handling dynamic columns. It is available only within an IBM Informix ESQL/C program. See also descriptor and system-descriptor area.

sqlhosts

A file that identifies the types of connections the database server supports.

stack operator

Operators that allow programs to manipulate values that are on the stack.

staging-area blobspace

(Not for Extended Parallel Server) The blobspace where a database server temporarily stores TEXT or BYTE data that is being outmigrated to an optical storage subsystem.

statement

A line or set of lines of program code that describes a single action (for example, a SELECT statement or an UPDATE statement).

statement block

A unit of SPL program code that performs a particular task and is usually marked by the keywords begin and end. The statement block of an SPL routine is the smallest scope of reference for program variables.

statement identifier

An embedded variable name or SQL statement identifier that represents a data structure defined in a PREPARE statement. It is used for dynamic SQL statement management by Informix SQL APIs.

static table

A nonlogging, read-only permanent table.

status variable

A program variable that indicates the status of some aspect of program execution. Status variables often store error numbers or act as flags to indicate that an error has occurred.

storage space

A dbspace, blobspace, or sbspace that is used to hold data.

stored procedure

A legacy term for an SPL routine.

Stored Procedure Language (SPL)

An Informix extension to SQL that provides flow-control features such as sequencing, branching, and looping. See also SPL routine.

strategy function

See operator-class strategy function.

string

A sequence of characters (typically alphanumeric) that is manipulated as a single unit. A string might consist of a word (such as 'Smith'), a set of digits representing a number (such as '19543'), or any other collection of characters. Strings generally are delimited by single quotes. String is also a character data type, available in IBM Informix ESQL/C programs, in which the character string is stripped of trailing blanks and is NULL terminated.

subordinate server

Any database server in a distributed query that did not initiate the query. Sometimes called remote server. See also coordinating server.

subordinate table

See outer join.

subquery

A query that is embedded as part of another SQL statement. For example, an INSERT statement can contain a subquery in which a SELECT statement supplies the inserted values in place of a VALUES clause; an UPDATE statement can contain a subquery in which a SELECT statement supplies the updating values; or a SELECT statement can contain a subquery in which a second SELECT statement supplies the qualifying conditions of a WHERE clause for the first SELECT statement. (Parentheses always delimit a subquery, unless you are referring to a CREATE VIEW statement or unions.) Subqueries are always parallelized. See also correlated subquery and independent subquery.

subscript

A subscript is an integer-valued offset into an array. Subscripts can be used to indicate the start or end position in a character data-type variable.

substring

A portion of a character string.

subtable

A typed table that inherits properties (column definitions, constraints, triggers) from a supertable above it in the table hierarchy and can add additional properties.

subtype

A named row type that inherits all representation (data fields) and behavior (routines) from a supertype above it in the type hierarchy and can add additional fields and routines. The number of fields in a subtype is always greater than or equal to the number of fields in its supertype.

supertable

A typed table whose properties (constraints, storage options, triggers) are inherited by a subtable beneath it in the table hierarchy. The scope of a query on a supertable is the supertable and its subtables.

supertype

A named row type whose representation (data fields) and behavior (routines) is inherited by a subtype below it in the type hierarchy.

support function

See aggregate support function, opaque-type support function, and operator-class support function.

support routine

See support function.

symmetric multiprocessing system

A system composed of multiple computers that are connected to a single high-speed communication subsystem. An SMP has fewer computers than an MPP system and cannot be partitioned into nodes. Compare with massively parallel processing system.

synonym

A name that is assigned to a table, view, or sequence and that can be used in place of the original name. A synonym does not replace the original name; instead, it acts as an alias for the table, view, or sequence.

sysmaster database

A database on each database server that holds the ON–Archive catalog tables and system-monitoring interface (SMI) tables that contain information about the state of the database server. The database server creates the sysmaster database when it initializes disk space.

system call

A routine in an operating-system library that programs call to obtain information from the operating system.

system catalog

A group of database tables that contain information about the database itself, such as the names of tables or columns in the database, the number of rows in a table, the information about indexes and database privileges, and so on. See also data dictionary.

system-defined cast

A pre-defined cast that is known to the database server. Each built-in cast performs automatic conversion between two different built-in data types.

system-defined hash fragmentation

An Extended Parallel Server-defined distribution scheme that maps each row in a table to a set of integers and uses a system-defined algorithm to distribute data evenly by hashing a specified key.

system-descriptor area

A dynamic SQL management structure that is used with the ALLOCATE DESCRIPTOR, DEALLOCATE DESCRIPTOR, DESCRIBE, GET DESCRIPTOR, and SET DESCRIPTOR statements to store information about database columns or host variables used in dynamic SQL statements. The structure contains an item descriptor for each column; each item descriptor provides information such as the name, data type, length, scale, and precision of the column. The system-descriptor area is the X/Open standard for handling dynamic columns. See also descriptor and sqlda.

system index

An index that the database server creates to implement a unique constraint or a referential constraint. A system index is distinct from a user index, which a user creates explicitly.

system-monitoring interface

A collection of tables and pseudo-tables in the sysmaster database that maintains dynamically updated information about the operation of the database server. The tables are constructed in memory but are not recorded on disk. Users can query the SMI tables with the SELECT statement of SQL.

table

A rectangular array of data in which each row describes a single entity and each column contains the values for each category of description. For example, a table can contain the names and addresses of customers. Each row corresponds to a different customer and the columns correspond to the name and address items. A table is sometimes referred to as a base table to distinguish it from the views, indexes, and other objects defined on the underlying table or associated with it.

table fragment

Zero or more rows that are grouped together and stored in a dbspace that you specify when you create the fragment. A virtual table fragment might reside in an sbspace or an extspace.

table fragmentation

A method of separating a table into potentially balanced fragments to distribute the workload and optimize the efficiency of the database operations. Also known as data partitioning. Table-fragmentation methods (also known as distribution schemes) include expression-based, hybrid, range, round-robin, and system-defined hash.

table hierarchy

A relationship you can define among typed tables in which subtables inherit the behavior (constraints, triggers, storage options) from supertables. Subtables can add additional constraint definitions, storage options, and triggers.

table inheritance

The property that allows a typed table to inherit the behavior (constraints, storage options, triggers) from a typed table above it in the table hierarchy.

target table

The underlying base table that a violations table and diagnostics table are associated with. You use the START VIOLATIONS TABLE statement to create the association between the target table and the violations and diagnostics tables.

tblspace

The logical collection of extents that are assigned to a table. It contains all the disk space that is allocated to a given table or table fragment and includes pages allocated to data and to indexes, pages that store TEXT or BYTE data in the dbspace, and bitmap pages that track page use within the extents.

TCP/IP

The specific name of a particular standard transport layer protocol (TCP) and network layer protocol (IP). A popular network protocol used in DOS, UNIX, and other environments.

temporary

An attribute of any file, index, or table that is used only during a single session. Temporary files or resources are typically removed or freed when program execution terminates or an online session ends.

terabyte

A unit of storage, equal to 1024 gigabytes or 1024⁴ bytes.

TEXT

A data type for a simple large object that stores text and can be as large as 2³¹ bytes. See also BYTE.

thread

A piece of work or task for a virtual processor just as a virtual processor is a task for a CPU. A virtual processor is a task that the operating system schedules for execution on the CPU; a database server thread is a task that a virtual processor schedules internally for processing. Threads are sometimes called lightweight processes because they are like processes but make fewer demands on the operating system. See also multithreading and user thread.

TLI

Acronym for Transport Layer Interface. It is the interface designed for use by application programs that are independent of a network protocol.

trace

To keep a running list of the values of program variables, arguments, expressions, and so on, in a program or SPL routine.

transaction

A collection of one or more SQL statements that is treated as a single unit of work. If one statement in a transaction fails, the entire transaction can be rolled back (canceled). If the transaction is successful, the work is committed and all changes to the database from the transaction are accepted. See also explicit transaction, implicit transaction, and singleton implicit transaction.

transaction lock

A lock on an R-tree index that is obtained at the beginning of a transaction and held until the end of the transaction.

transaction logging

The process of keeping records of transactions. See also logical log.

transaction mode

The method by which constraints are checked during transactions. You use the SET statement to specify whether constraints are checked at the end of each data manipulation statement or after the transaction is committed.

trigger

A database object that executes a set of actions if a DML event manipulates a specified table. (An INSTEAD OF trigger substitutes a set of actions for a DML event that attempts to manipulate a specified view.)

tuple

See row.

two-phase commit

A protocol that ensures that transactions are uniformly committed or rolled back across multiple database servers. It governs the order in which commit transactions are performed and provides a recovery mechanism in case a transaction does not execute. See also heterogeneous commit.

type constructor

An SQL keyword that indicates to the database server the type of complex data to create (for example, LIST, MULTISET, ROW, SET).

type hierarchy

A relationship that you define among named row types in which subtypes inherit representation (data fields) and behavior (routines) from supertypes and can add more fields and routines.

type inheritance

The property that allows a named row type or typed table to inherit representation (data fields, columns) and behavior (routines, operators, rules) from a named row type above it in the type hierarchy.

type substitutability

The ability to use an instance of a subtype when an instance of its supertype is expected.

typed collection variable

An ESQL/C collection variable or SPL variable that has a defined collection data type associated with it and can only hold a collection of its defined type. See also untyped collection variable.

typed table

A table that is constructed from a named row type and whose rows contain instances of that row type. A typed table can be used as part of a table hierarchy. The columns of a typed table correspond to the fields of the named row type.

UDA

See user-defined aggregate.

UDF

See user-defined function.

UDR

See user-defined routine.

UDT

See user-defined data type.

unbuffered disk I/O

Disk I/O that is controlled directly by the database server instead of the operating system. This direct control helps improve performance and reliability for updates to data. Unbuffered I/O is supported by character-special files on UNIX and by both unbuffered files and the raw disk interface on Windows.

Uncommitted Read

See Read Uncommitted.

uncorrelated subquery

See independent subquery.

unique constraint

Specifies that each entry in a column or set of columns has a unique value.

unique index

An index that prevents duplicate values in the indexed column.

unique key

See primary key.

UNIX real user ID

See Informix user ID.

unload job

The information required to unload data from a relational database using the HPL. This information includes format, map, query, device array, project, and special options.

unlock

To free an object (database, table, page, or row) that has been locked. For example, a locked table prevents others from adding, removing, updating, or (in the case of an exclusive lock) viewing rows in that table as long as it is locked. When the user or program unlocks the table, others are permitted access again.

unnamed row type

A row type created with the ROW constructor that has no defined name and no inheritance properties. Two unnamed row types are equivalent if they have the same number of fields and if corresponding fields have the same data type, even if the fields have different names.

untyped collection variable

A generic ESQL/C collection variable or SPL variable that can hold a collection of any collection data type and takes on the data type of the last collection assigned to it. See also typed collection variable.

updatable view

A view whose underlying table can be modified by inserting values into the view. Only an INSTEAD OF trigger can update a multi-table view.

update

The process of changing the contents of one or more columns in one or more existing rows of a table.

update lock

A promotable lock that is acquired during a SELECT...FOR UPDATE. An update lock behaves like a shared lock until the update actually occurs, and it then becomes an exclusive lock. It differs from a shared lock in that only one update lock can be acquired on an object at a time.

user-defined aggregate

An aggregate function that is not provided by the database server (built in) that includes extensions to built-in aggregates and newly defined aggregates. The database server manages all aggregates.

user-defined base type

See opaque data type.

user-defined cast

A cast that a user creates with the CREATE CAST statement. A user-defined cast typically requires a cast function. A user-defined cast can be an explicit cast or an implicit cast.

user-defined data type

A data type that you define for use in a relational database. You can define opaque data types and distinct data types.

user-defined function

A user-defined routine that returns at least one value. You can write a user-defined function in SPL (SPL function) or in an external language that the database server supports (external function).

user-defined procedure

A user-defined routine that does not return a value. You can write a user-defined procedure in SPL (SPL procedure) or in an external language that the database server supports (external procedure).

user-defined routine

A routine that you write and register in the system catalog tables of a database, and that an SQL statement or another routine can invoke. You can write a user-defined routine in SPL (SPL routine) or in an external language (external routine) that the database server supports.

user ID

Also called authorization identifier. See also Informix user ID.

user ID password

See Informix user password.

user index

An index that a user creates explicitly with the CREATE INDEX statement. Compare with system index.

user name

See Informix user ID.

user password

See Informix user password.

user thread

User threads include session threads (called sqlexec threads) that are the primary threads that the database server runs to service client applications. User threads also include a thread to service requests from the onmode utility, threads for recovery, and page-cleaner threads. See thread.

variable

The identifier for a location in memory that stores the value of a program object whose value can change during program execution. Compare with constant, macro, and pointer.

variant function

A user-defined function that might return different values when passed the same arguments. A variant function can contain SQL statements. Compare with nonvariant function.

view

A dynamically controlled subset of the columns of one or more database tables. A view can give the programmer control over what information the user sees and manipulates and represents a virtual table that holds the results of a specified SELECT statement.

violations table

A special table that holds rows that fail to satisfy constraints and unique index requirements during data manipulation operations on base tables. You use the START VIOLATIONS TABLE statement to create a violations table and associate it with a base table.

virtual column

A derived column of information, created with an SQL statement that is not stored in the database. For example, you can create virtual columns in a SELECT statement by arithmetically manipulating a single column, such as multiplying existing values by a constant, or by combining multiple columns, such as adding the values from two columns.

virtual-column index

A type of generalized-key index that contains keys that are the result of an expression.

virtual processor

A multithreaded process that makes up the database server and is similar to the hardware processors in the computer. It can serve multiple clients and, where necessary, run multiple threads to work in parallel for a single query.

virtual table

A table created to access data in an external file, external DBMS, smart large object, or in the result set of an iterator function in a query. The database server does not manage external data or directly manipulate data within a smart large object. The Virtual-Table Interface allows users to access the external data in a virtual table using SQL DML statements and join the external data with Dynamic Server table data.

VLDB

Acronym for very large database.

warning

A message or other indicator about a condition that software (such as the database server or compiler) detects. A condition that results in a warning does not necessarily affect the ability of the code to run. See also compile-time error and runtime error.

white space

A series of one or more space characters. The GLS locale defines the characters that are considered to be space characters. For example, both the TAB and blank might be defined as space characters in one locale, but certain combinations of the CTRL key and another character might be defined as space characters in a different locale.

wide character

A form of a code set that involves normalizing the size of each multibyte character so that each character is the same size. This size must be equal to or greater than the largest character that an operating system can support, and it must match the size of an integer data type that the C compiler can scale. Some examples of an integer data type that the C compiler can scale are short integer (short int), integer (int), or long integer (long int).

wildcard

A special symbol representing any character or any string of zero or more characters. In SQL, for example, you can use the asterisk ( * ), question mark ( ? ), percent sign ( % ), and underscore ( _ ) as wildcard characters in some contexts. (Asterisk and question mark are also UNIX wildcards.)

window

A rectangular area on the screen in which you can take actions without leaving the context of the background program.

WORM

Acronym for Write-Once-Read-Many optical media. When a bit of data is written to a WORM platter, a permanent mark is made on that optical platter.

X/Open

An independent consortium that produces and develops specifications and standards for open-systems products and technology, such as dynamic SQL.

X/Open Portability Guide (XPG)

A set of specifications that vendors and users can use to build portable software. Any vendor that carries the XPG brand on a given software product is guaranteeing that the software correctly implements the X/Open Common Applications Environment (CAE) specifications. There are CAE specifications for SQL, XA, ISAM, RDA, and so on.

zoned decimal

(1) A data representation that uses the low-order four bits of each byte to designate a decimal digit (0 through 9) and the high-order four bits to designate the sign of the digit.

(2)