In the os/390 world, emptying a table without any logging overheads was quite simple.
Simply LOAD REPLACE using an DUMMY dataset WITH LOG NO option.
Is there an equivalent concept for UDB Db2 on Windows.
I tried the following but no luck.
1. EXPORT from tab1 (making sure no rows are returned)
use the LOAD command and specify the empty file.
LOAD didn't go thru because it didn't like the file format
2. EXPORT 1 row from tab1
use LOAD with REPLACE option and non-recoverable (for LOG NO)
one row was replaced, which means row count hasn't changed. the purpose of loading one row (replacing data with one row) was defeated.
Bottomline, my question is
In UDB DB2 for Windows, how do you empty a large table without incurring any LOG overheads ?