comparing

[bobby]

hi,
i am using compare of columns for finding delta ,like if linka<>hashlinkb
then linka else hashlinkb and applying constriant on hashfilenotfound
Now I have some problems
1) still duplicates rows do pass
2)error comes like hash file is null.
(hash file is from same target ,if load is incremented hash file seems to be not incremented)
plz suggest,
Thanks
Bobby

[kcbland]

First, you're posting in the wrong forum. Server questions are for the Server forum, this forum is for Parallel jobs.

Second, I'm trying to follow your wording, but it's confusing. If you're writing to a hash file and referencing it as well, you have to NOT use any read or write caching, as reads will not find previous writes if you're within the write-delay cache block (it hasn't flushed yet). So, make sure you're not caching.

Third, you need to describe your job design, meaning stages and links. Something like:

I'm reading from a sequential file into a transformer, where I'm referencing a hash file. If the reference lookup is not found, I output to an ODBC stage using INSERT ONLY SQL and also write back to the same reference hash file. If the reference lookup is found, then I output to a different ODBC stage using UPDATE EXISTING SQL.

This is a friendly hint because your posts are difficult to follow. The Inner Circle crowd have almost 10000 posts between them, we are in a position to nicely ask that questors put some care into their questions.

Thanks!