DSXchange

It was not clear from your description whether the hash partitioning was based on precisely the same key column(s) in all jobs, particularly in Job 7 and Job 8. If this is in fact the case then you ought not to get incorrect results. On that basis, I suspect that somewhere the partitioning keys are ...

NEVER kill DataStage processes.

Which part of that was unclear?

Get someone who does have access to the server to forward a copy of the log file to you, or get (read-only) access to the log file on the server.

Aaarrgh!! The extra expense!!!

There is almost certainly an ODBC driver out there. jBase is one of the family of databases, like UniVerse and Unidata, the support "nested relational" designs - multi-valued fields and so on.

bmouton wrote:They are still HASH files

No they're not. They are HASHED files.

A hash file is a tool used for shaping a block of hash. Whether that means ground meat or something else to you is immaterial.

Without knowing more about your servers, particularly their performance data (CPU speed and the like) it's really impossible to say. But nearly 40,000 rows per second sustained over 90 minutes would keep most people here happy.

is an alternative; it has the advantage that it will remove currency symbols as well.

In version 8.1 you get the choice of where the logs are. It's my understanding that the default is to put them in the unified metadata repository (that is, the XMETA database if you didn't select a different name).

Make very sure too that the four environment variables for compiler and linker are correctly set for the new platform, and that 32-bit libraries are found ahead of 64-bit libraries in every instance.

There's nothing you can tune with Hash aggregation method. Sort will give better performance (measured as elapsed time) but does require sorted input. Use an explicit Sort stage and use as much memory for sorting as you can afford. Don't forget to partition and sort on the grouping keys - do this as...

There's nothing you can tune with Hash aggregation method. Sort will give better performance (measured as elapsed time) but does require sorted input. Use an explicit Sort stage and use as much memory for sorting as you can afford. Don't forget to partition and sort on the grouping keys - do this as...

It's not a bug. The problem is between the keyboard and the chair.

And, by the way, a Sequencer does not have any expressions. It only has one property, and that only takes two possible values - "Any" or "All".

Please be careful with terminology.

Well spotted! We could have spent ages without guessing about the mass change!

And a valuable lesson about the dangers of unintended changes during global search and replace.