DSXchange

The operating system will have a limit on the total length of a command string. DataStage imposes no limit, but is constrained by the operating system limit.

There are some RPMs that have to be installed on Linux. These are listed in the on-line documentation (Information Center) and referred to in the Planning, Installation and Configuration Guide.

OK, you know to look at SystemOut.log. That's a good start. Is there anything in there that you understand (in a diagnostic sense)?

Did you also look in SystemErr.log?

You may get "duplicate hostname on network" errors.

DataStage might work on the independent machines, or it might not because of this. It is required that Information Server components be on a machine that can resolve its hostname to a unique IP address.

MSXML is required if you are using any metabrokers or bridges.

I have not yet used Metadata Asset Manager - I'm still in the habit of using istool for imports. Perhaps you could try that as an alternative.

Can you see the reports on the Reporting tab of Web console for Information Server?

What James said. Hash aggregation method instructs the Aggregator stage to build a hash table in memory (per partition) in which to store results. Nothing comes out of such an Aggregator stage until all input rows have been consumed, because there is no a priori knowledge about which is the last row...

vmcburney wrote:Informatica have blatantly copied the IBM Business Glossary in look and feel so I don't know how they compare.

There's not really any other way you can present a list of terms, their definitions and other characteristics!

You really need to use Hash partitioning algorithm, specifying your grouping column as the partitioning key. What's happening under Auto is round-robin partitioning, so that your grouping value might occur on either or both of your two nodes. If you were running four nodes you might see quadruplicat...

How are your data partitioned?

Users can be deleted from the internal registry irrespective of which user registry is in use, and even when WAS is shut down, using DirectoryAdmin.bat -delete_users command (or DirectoryAdmin.sh -delete_users on UNIX),

A couple of points.

The dscmdexport command is only available on the client. Therefore it can not be used in a UNIX script.

If you're exporting the whole project (*/*.*) then -includedependent is redundant.

Put simply, your data and metadata don't match.

Add a reject link to capture the rejected records. Compare these to the metadata that describe them.

Apologies - DataStage is looking for the record delimiter, not the field delimiter. Try setting the record delimiter to "none".

Specify "none" as the field delimiter.