Load dump file

Specifying the dumpfile file type modifier tells the load utility the name and the location of the exception file to which rejected rows are written.

When running in a partitioned database environment, rows can be rejected either by the partitioning subagents or by the loading subagents. Because of this, the dump file name is given an extension that identifies the subagent type, as well as the database partition number where the exceptions were generated. For example, if you specified the following dump file value:

   dumpfile = "/u/usrname/dumpit"

Then rows that are rejected by the load subagent on database partition five will be stored in a file named /u/usrname/dumpit.load.005, rows that are rejected by the load Subagent on database partition two will be stored in a file named /u/usrname/dumpit.load.002, and rows that are rejected by the partitioning subagent on database partition two will be stored in a file named /u/usrname/dumpit.part.002, and so on.

For rows rejected by the load subagent, if the row is less than 32 768 bytes in length, the record is copied to the dump file in its entirety; if it is longer, a row fragment (including the final bytes of the record) is written to the file.

For rows rejected by the partitioning subagent, the entire row is copied to the dump file regardless of the record size.