DSXchange

If you write a quickie job that does a lookup on your table for all the rows you can get a list of records that weren't inserted (the lookup will fail for those rows). Can you discern any pattern in those record? If you re-run the whole thing again to the same records not get inserted? What happens ...

sun rays : Thanks - I didn't even think of the duplicates; I just jumped to the assumption of rejected rows!

No need to guess - some of the rows are being rejected by the database. Depending upon the database and stage type you are using you have various ways of capturing these rejects. Please look up the appropriate stage type in your online documentation and read up on how to implement error handling and...

The syntax is SQL, but there was a thread here earlier today which will answer your question.

Your file has 189 bytes in the first record, and your column metadata only account for 188 of those bytes. You are using a fixed width format, so the actual length must match that of your column definitions. What is are your column definitions? You could use an editor to manually check to see where ...

but the options are numeric i can't able to give uv there.the opitons start like 0,1,2 etc any more things do i need to do Thanks, RK I don't quite understand. If you telnet to the Windows server, it will prompt you for the name of the account to connect to. If you call up the uvsh or dssh program ...

DataStage export itself unfortunately will not do this. I don't know if any of the other tools such as Compilewitch do this dependency checking, so you might still have a chance to get this automated. I usually end up exporting the jobs one-by-one to the same .dsx file; but it is tedious and error-p...

Although the option of writing an external call is open to you (I think that you can use the IBM documents for the UCI for that), I think you'll be much happier of integrating a call to MyEnv = DSExecute('UNIX','env') and then parsing the returned values in that string. The line-feeds in the...

The RT_SC file is filled at compile with the scripts that PX uses - does the error go away if you recompile the job?

Since DataStage uses the host system's scheduler and you are on UNIX your system is most likely CRON and you can modify the entry in the crontab yourself; I couldn't find a specifi date option in the Director GUI to do so.

Sheshadri, is your rule to do this one character at a time? I can't think of a one-liner to do this, but a very simple user-written routine would do it. MyRoutine(ColA,ColB) EQUATE LF TO CHAR(10) EQUATE Separator TO ',' ColBLength = LEN(ColB) Ans = '' FOR Index = 1 TO ColBLength Ans ...

ian, it has a default value without quotes, since that is the value it is using when you start your job manually. But when called through your scheduled process it is getting the extra double-quotes from somewhere; this is what you need to analyze. Most likely the double-quotes were added to the par...

So the extra quotes are being added when you run through the scheduler... How is the scheduler getting/setting that value?

The analysis tools require that a file has a local VOC entry. You can use the "SETFILE /path/to/your/file DUMMYFILENAME" and then a "ANALYZE.FILE DUMMYFILENAME" to do this.

Break it up into 4 or 5 lookups per transformer. This is assuming you have a multi-cpu system; otherwise none of this makes an effective difference.