Administration Guide

Hints and Tips

1. Test with small amounts of data to get familiar with the tool before loading huge amounts.
2. If the input data is already sorted (or in some order) and you wish to maintain that order during the load process, then use only one database partition for splitting. Parallel splitting will not guarantee that the data will be loaded in the same order as it was received.
3. If LOBs are stored within separate files (that is, you are making use of the LOBSINFILE feature), then all directories containing LOB files should be made accessible to all the database partitions where loading is taking place.
4. AutoLoader will ignore the MESSAGES parameter in a LOAD command, and it will direct all messages from the LOAD command into the file load_log.XXX as described in "Usage" above.
5. AutoLoader will only choose one output database partition to collect statistics. The option RUN_STAT_NODE in the AutoLoader configuration file is provided for this purpose.
6. Multiple invocations of AutoLoader may be used to load data simultaneously into separate tables. However, make sure that:
   1. The tables reside on separate table spaces.
   2. All the AutoLoader's are invoked from separate directories.
   3. AutoLoader uses the datafile name to create some temporary pipes. Ensure that the file name is unique for the multiple AutoLoader's.

[ DB2 List of Books | Search the DB2 Books ]