SYSDISTRIB

Home | Previous Page | Next Page System Catalog Tables >

SYSDISTRIB

The sysdistrib system catalog table stores data-distribution information for the database server to use. Data distributions provide detailed table and column information to the optimizer to improve the choice of execution paths of SELECT statements. The sysdistrib table has the following columns.

Column	Type		Explanation
tabid	INTEGER		Code uniquely identifying the table where data values were gathered
colno	SMALLINT		Column number in the source table
seqno	INTEGER		Ordinal number for multiple entries
constructed	DATE		Date when the data distribution was created
mode	CHAR(1)		Optimization level: M = Medium H = High
resolution	SMALLFLOAT		Specified in the UPDATE STATISTICS statement
confidence	SMALLFLOAT		Specified in the UPDATE STATISTICS statement
encdat	STAT CHAR(256)	IDS XPS	Statistics information
type (IDS)	CHAR(1)		Type of statistics: A = encdat has ASCII-encoded histogram in fixed-length character field S = encdat has user-defined statistics

Information is stored in the sysdistrib table when an UPDATE STATISTICS statement with mode MEDIUM or HIGH is executed for a table. (UPDATE STATISTICS LOW does not insert a value into the mode column.)

Only user informix can select the encdat column.

Each row in the sysdistrib system catalog table is keyed by the tabid and colno for which the statistics are collected.

For built-in data type columns, the type field is set to A. The encdat column stores an ASCII-encoded histogram that is broken down into multiple rows, each of which contains 256 bytes.

In Dynamic Server, for columns of user-defined data types, the type field is set to S. The encdat column stores the statistics collected by the statcollect user-defined routine in multirepresentational form. Only one row is stored for each tabid and colno pair. A composite index on the tabid, colno, and seqno columns requires unique combinations of values.