DSXchange

Even a simple "** This is a comment line" in the job control is sufficient.

Hmm.. so far the file is still below the 2Gb limit that would necessitate using a 64bit file (which is slower than the default file). Can you output the "ANALYZE.FILE m_HFg2_6 STATS" output - that would give some more information. Do you know what the keys look like and if they are more (o...

I would suggest doing an export to a .dsx file, then using a global replace to put in the new strings.

Can you post your actual stage variable derivations, please?

All objects at Version 7 exist and will run at version 8; but the metadata repository at Version 8 and beyond is now separated from the old server engine and hashed files. In addition, user and process management is done within the web console of IIS. You will need to read up on this new functionali...

Are you outputting from the transform stage straight to a sequential file? This method will only work if you do it that way.

Normally the migration is straight-forward and direct. The problem arises when you have a production environment and cannot afford to discover that you aren't part of the 99% that have no issues during the upgrade. At Version 8 there are some fundamental changes to the product and whoever is respons...

"col1" is not valid, you need to use {linkname}.col1; likewise the assign statement needs you to replace "{NewValue}" with your linkname.col1.

To add to what Ray has already stated, use the Microsoft "msquery" tool to check that your ODBC connection is set up correctly. Once you get msquery to work you will almost certainly also have the DataStage connectivity working.

That is a <CR><LF> sequence for DOS file format. Even though that line only writes 1 row to the output, it creates 2 rows when that file is read again from the sequential file due to the added carriage return / line feed.

Is your DataStage server on Windows or UNIX? If on UNIX, then you will need a UNIX ODBC driver to Excel, creating a Windows system DSN will only work with a Windows DataStage server. I think for .xls ODBC connections you don't need a userid or password.

umm..

I just did tell you my idea, using a simple "ps -ef" on a userid is not helpful in large sites.

I have a couple of PX jobs here which fire up 400-500 UNIX pids apiece, which is why I prefer to get the "parent" than list all the players.

In that case your memory requirements for this stage are higher than what the machine can give you. Since you cannot increase your hard data segment size you will need to reduce your storage requirements, which can be done in this case by sorting your data before the aggregator stage.