DSXchange

Your logic is wrong, which is why things are repeating.

The simple solution would be to read each line as a single VarChar and to use a Convert() function to remove the double-quote characters.

What would happen if you were doing this in a database?

Would you expect DataStage to behave the same way?

Check the "do not checkpoint run" property in the Routine activity to prevent it from being checkpointed (to force the routine to be run every time the sequence runs).

I was just wondering if it is possible to define the variable at project level and enter the value at run level. So that parameter name is same across all the command tasks. It is not possible. This might be a case for creating the job control routine manually (perhaps based on the code generated b...

Add a path to the sequence such that it can finish immediately the job has finished.

DataStage does it the same way database servers do it: pairwise from the left. So, if you have three inputs for example, the left and intermediate inputs are joined then the result of that is joined with the right input.

You can specify whatever values you like for the values, using options Copy Code, Delete Code, Edit Code and Insert Code. Default values for these are 0, 2, 3 and 1 respectively. As you select each of these properties, the Information pane defines what it means. For example, a copy means that "...

Wait till the weekend is over.

It won't do any harm to contact editor@dsxchange.net, but these people actually have a life as well as running DSXchange.

There isn't directly, as that would be anathema to the concept of "surrogate key".

You could do it, but not within the context of a single job.

Define what you mean by "performance" in this context. Certainly generating frequencies will generate more accurate results (for a large enough sample) than an artificially flat frequency distribution.

As a general rule, a native Connector should be preferred over ODBC.

For DB2 there's not a huge amount of difference, as the DB2 API is modelled on ODBC standards. But there is one major difference; the DB2 Connector stage inherently handles DB2 partitioning.

I prefer not to use RCP in general, though I do use it occasionally when some kind of generic solution is required. Some of my reluctance stems from earlier releases, when RCP was not well supported in lineage reports. It can also lead to hard-to-diagnose "column has been dropped" warnings...

Welcome aboard.

What error message(s) occur in the sequence's log when this error occurs?

What exit status was returned to, and processed by, Control-M in this case?

Please post exact error messages.

The Lookup stage permits return of multiple rows from one reference link.

Open the Lookup Stage Conditions and choose the reference link name from the "Multiple rows returned from link" drop-down list.

This is probably most easily accomplished by sorting the data in reverse (or reading the file backwards), adding the "headers" at the end of each group, then sorting again back into forward order.