For details about the tables that are subject to extraction processing, see 1.3 Tables supported for data linkage.

For details about the data types that are subject to extraction processing, see 1.4.1 Data types supported by Datareplicator.

For details about the correspondence between the source and target databases, see 4.3.4 Designing for the supported data types.

For details about the SQL statements that are subject to extraction processing, see 1.5 SQL statements supported for data linkage.

The source Datareplicator executes data linkage on the basis of the database update log in the system log. Therefore, all update processing on a database subject to data linkage by a UAP that outputs database update log information is subject to extraction processing.

When you execute a UAP to update a database that is to subject to data linkage while data linkage is in effect, some environment variables and SQL statements require special handling. The following table provides notes on UAP execution.

Table 4-4 Notes on UAP execution

Item	Notes
Specification of no-log mode in an environment variable	Do not specify NO in the PDDBLOG environment variable in HiRDB's client environment definition.
Modification of definitions by SQL statements	While data linkage is in effect, do not issue any of the following SQL statements for a database that is subject to data linkage: PURGE TABLE (for a partitioned table spanning multiple back-end servers#) ALTER TABLE DROP TABLE For details about how to modify definitions during execution of data linkage, see 6.5.6 Source HiRDB handling procedure.

#: When the source Datareplicator detects a PURGE TABLE statement for a table that spans multiple back-end servers, it outputs a warning message to the error information file without extracting the corresponding update information. If this happens, you can use the prg_eventno operand in the transmission environment definition to terminate transmission processing or to issue an event to the target Datareplicator.

HiRDB Datareplicator supports only the database load utility (pdload) for extraction. Note that if there is an error in specifying options, inconsistency might occur between extraction and import processing. Note also that limitations in Datareplicator preclude its use to extract BLOB data.

If other utilities that output update log information are executed during data linkage, inconsistency occurs between extraction and import processing. To execute other utilities, you must specify the options so as to prevent them from operating within the extraction range.

The following table provides notes on utility execution.

Table 4-5 Notes on utility execution

Utility name		Notes
Database definition utility	pddef	While data linkage is in effect, do not issue any of the following SQL statements for a database that is subject to data linkage: ALTER TABLE DROP TABLE For details about how to modify the definitions during execution of data linkage, see 6.5.6 Source HiRDB handling procedure.
Database load utility	pdload	Existing table Do not execute the utility in the creation mode during execution of data linkage (you cannot specify the -d option). Specify the log acquisition mode in the addition mode (specify a in the -l option). New table To add a new table subject to linkage during execution of data linkage, use the log acquisition mode (specify a in the -l option). If you will be applying data linkage to pdload, also see the chapter that includes 4.7.3(9)(a) Number of locked resources.
Database structure modification utility	pdmod	During execution of data linkage, do not re-initialize an RDAREA that contains a table subject to extraction processing.
Database reorganization utility	pdrorg	During execution of data linkage, use the no-log mode or pre-update log acquisition mode (specify n or p in the -l option). If you use the log acquisition mode, conformity is lost between the source and target databases because update information about only a portion of the reorganized database is output to the system log.
Database recovery utility	pdrstr	To recover a database, use the database's backup copy and the log collected since the backup was made. If you attempt to recover a database by any other method, conformity is lost between the source and target databases.
Rebalancing utility	pdrbal	To execute the rebalancing utility on a table subject to data linkage, you must modify the table definition (addition of RDAREAs). For details about how to modify a table definition and an extraction definition, see (3) in 6.5.6 Source HiRDB handling procedure. Database update processing by the rebalancing utility is not subject to extraction processing.

Note:

If a table subject to utility processing contains a BLOB column, the rows updated by the utility will not be imported into the target database because the source Datareplicator ignores database updating by a utility.

The maximum amount of update information that can be extracted by the source Datareplicator (sum of the definition lengths of columns subject to extraction) is 256 megabytes.

If the amount of update information exceeds 256 megabytes, the extraction definition preprocessor command results in an error. Note that SGMLTEXT-type and XML-type columns are not subject to checking. To extract SGMLTEXT-type and XML-type columns, you must take into account the data lengths handled as these types so that the columns to be extracted do not exceed 256 megabytes.