Recent content by DBGod

Not sure what the signature pads give as output... If they produce an image file, then use a LOB type. If they produce binary data, then maybe CHAR() FOR BIT DATA. Regards Mick

The short answer is, it all depends! If there is little variation in the column lengths, the stick with CHAR. If there's lots of difference in lengths, use VARCHAR for maximum lengths over 30 say. On uncompressed tables, updates of VARCHAR-containing rows can lead to rows being moved to other...

This is an indication that the load module and package did not originate from the same precompile step. Check:- Did you bind the correct DBRM member? Have you bound in the correct DB2 sub-system? Are you running with the correct load library? You should check each step of the compile/bind...

Don't know much about the Unix side....I'm more OS/390. But:- 1) Execution times between SPs and queries should be similar if the same access path is being used. Stored Procedures are generally very reliable, and provide a method of standardizing database accesses and also reducing network...

If there is an index on STOREID, DB2 may or may not use this to perform the DISTINCT action, depending on the stats for the index held in the catalog. If the index is not used, then DB2 will execute a tablespace scan, followed by a sort and elemination of duplicate rows. If there are many rows...

What platform are you executing on? Mick

Try something like:- SELECT SUBSTR(CHAR(timestamp),1,10) CONCAT SUBSTR(CHAR(timestamp),12,15) FROM table_name Hope that helps Mick