IBM Books

API Reference


Long Field Manager Log Records

Long field manager log records are written only if a database is configured with LOG RETAIN on or USEREXITS enabled; They are written whenever long field data is inserted into a table. When long field data is updated, the update is treated as a delete of the old long field value, followed by an insert of the new value.

To conserve log space, long field data inserted into tables is not logged if the database is configured for circular logging. In addition, when a long field value is updated, the before image is shadowed and not logged.

All long field manager log records begin with a header.

All long field manager log record offsets are from the end of the log manager log record header.

Table 101. Long Field Manager Log Record Header (LongFieldLogRecordHeader)
Description Type Offset (Bytes)
Originator code (component identifier = 3) unsigned char 0(1)
Operation type unsigned char 1(1)
Pool identifier unsigned short 2(2)
Object identifier unsigned short 4(2)
Parent pool identifiera unsigned short 6(2)
Parent object identifierb unsigned short 8(2)
Total Length: 10 bytes
Note:
a Pool ID of the data object
b Object ID of the data object

Add Long Field Record

This log record is written whenever long field data is inserted. The length of the data is rounded up to the next 512-byte boundary.

Table 102. Add Long Field Record Log Record Structure
Description Type Offset (Bytes)
Log header LongFieldLogRecordHeader 0(10)
Long field lengtha unsigned short 10(2)
File offsetb unsigned long 12(4)
Long field data char[ ] 16(variable)
Note:
a Long field data length in 512-byte sectors (actual data length is not logged)
b 512-byte sector offset into long field object where data is to be inserted


[ Top of Page | Previous Page | Next Page | Table of Contents | Index ]

[ DB2 List of Books | Search the DB2 Books ]