Fixed-Point Binary Representations

The DataBlade API provides the following data types to support the binary representations of SQL fixed-point data types.

Both the DECIMAL and MONEY data types use the same internal format to store a fixed-point value. For more information on this format, see Internal Fixed-Point Decimal Format.

DECIMAL Data Type: Fixed-Point Data

When you define a column with the DECIMAL(p,s) data type, the syntax of this definition specifies a fixed-point value for the column. This value has a total of p (<= 32) significant digits (the precision) and s (<= p) digits to the right of the decimal point (the scale).

The SQL DECIMAL data type holds the internal (binary) format of a decimal value. This value is a computer-independent method that represents numbers of up to 32 significant digits, with valid values in the range 10^-129 to 10⁺¹²⁵. For more information, see Internal Fixed-Point Decimal Format.

The DataBlade API supports the SQL DECIMAL data type with the mi_decimal data type. Therefore, the mi_decimal data type also holds the binary representation of a decimal value. The mi_numeric data type is a synonym for mi_decimal.

MONEY Data Type

When you define a column with the MONEY(p) data type, it has a total of p (<= 32) significant digits (the precision) and a scale of 2 digits.

You can also specify a scale with the MONEY(p,s) syntax, where s represents the scale. For a complete description of the MONEY data type, see the IBM Informix: Guide to SQL Reference.

The DataBlade API supports the SQL MONEY data type with the mi_money data type. The mi_money data type holds the internal (binary) format of a MONEY value. This binary representation of the MONEY data type has the same structure as the fixed-point DECIMAL data type. For more information, see Internal Fixed-Point Decimal Format.

The decimal.h Header File

The decimal.h header file contains definitions for use with the DECIMAL and MONEY data types. This header file defines the following items:

• The dec_t typedef
• The decimal macros
• The DECIMAL-type functions of the ESQL/C library

The mitypes.h header file automatically includes decimal.h. In turn, the milib.h header file automatically includes mitypes.h, and mi.h automatically includes milib.h. Therefore, you automatically have access to the dec_t structure, the mi_decimal and mi_money data types, any of the decimal macros, or any of the ESQL/C DECIMAL-type functions when you include mi.h in your DataBlade API module.

Internal Fixed-Point Decimal Format

The DECIMAL and MONEY data types store fixed-point values in an Informix-proprietary internal format: the dec_t structure. This structure holds the internal (binary) format of a DECIMAL or MONEY value, as follows:

#define DECSIZE 16

struct decimal
   {
   short dec_exp;
   short dec_pos;
   short dec_ndgts;
   char  dec_dgts[DECSIZE];
   };

typedef struct decimal dec_t;

This dec_t structure stores the number in pairs of digits. Each pair is a number in the range 00 to 99. (Therefore, you can think of a pair as a base-100 digit.) Table 14 shows the four parts of the dec_t structure.

Table 15 shows some sample dec_t values.

The mi_decimal and mi_money data types use the dec_t structure to hold the binary representation of a DECIMAL and MONEY value, respectively.

The Decimal Macros

The decimal.h header file also includes the following macros that might be useful in a DataBlade API module.

Decimal Macro

Description

DECLEN(p, s)

Calculates the minimum number of bytes required to hold the DECIMAL(p,s) value

DECPREC(size)

Calculates a default precision given the number of bytes (size) used to store the number

PRECTOT(dec)

Returns the total precision of the dec value

PRECDEC(dec)

Returns the scale of the dec value

PRECMAKE(p, s)

Creates a precision value from the specified total precision (p) and scale (s)

Tip:

For a complete list of decimal macros, consult the decimal.h header file that is installed with your database server. This header file resides in the incl/public subdirectory of the INFORMIXDIR directory.

ESQL/C DECIMAL-Type Functions

Because the binary representation of DECIMAL (mi_decimal) and MONEY (mi_money) values is an Informix-proprietary format, you cannot use standard system functions to perform decimal operations on mi_decimal and mi_money values. Instead, the DataBlade API provides support for the following ESQL/C functions on the DECIMAL and MONEY data types.

Any other operations, modifications, or analyses can produce unpredictable results.