DSXchange

When it aborted, did you reset it? If so, did a 'From previous run...' message show in the log?

Anything 'special' or different about your SQL in this case? Is it generated or user defined?

1) Add a transformer. If the problem persists,

2) Reset the aborted job and post any 'From previous run...' log entry.

shamik_CHAK wrote:if it is then why it is so roundly advised to do any aggregation using database utilities and not thru ds?

That advice relates to the Aggregator stage - why pull data out and do the 'aggregation' in DS when one of the supposed strengths of the database is its aggregation ability?

Also their The Data Warehouse ETL Toolkit is well worth owning. ISBN-13: 978-0764567575.

Can you approach this as a series of discrete steps? Evaluate your condition, possibly truncate, etc - all controlled in a Sequence job by triggers?

Welcome. First thing to note is that posting in ALL CAPS is considered SHOUTING. We use our inside voices here. Secondly, I haven't played with the Teradata stages. Does it have a option to 'Treat Warnings as Fatal'? That could be one way to force an abort. Another way might be to set a very low War...

Welcome. :D The sequential file is the easy part. You'll need to expand on your requirements, however. 'One row for one employee contains the most effective employee data' means what exactly? Of the 10 records, you need to pick only one? Or you need to combine information from the 10 records, pickin...

Welcome. :D 'Imported my backup' means what? You imported an export you made of a previous project? If so, what exactly did you export? Import? If this 'NB' function is missing from the project (verify via the Manager), you'll need to export it from wherever it does exist and import it into this pro...

This doesn't make much sense. Did you mean you want the output to 'look in the same pattern' as it is in the input XML? Is your output XML again or a flat file? Please clarify.

In any case, post your job design.

Ashish wrote:I 've two parallel jobs and both jobs append the same sequential file parallel do you thinkg so it work

No. Multiple processes cannot write to the same file at the same time, period. Even if both are in 'append' mode.

Since they seem to have caused the problem, suggest you go back to IBM for the solution. So, everyone has this problem with the Administrator now?

I never did find a way to cat files to standard out such that the Filter could read them in as one continuous stream. I just fell back on cat to a fixed name either before job or more typically in a specific separate 'cat' job so that I don't need to repeat it if the processing job needs to be rerun.

I would heavily question your 'need' to do this. Why not a job if you need to do Oracle work? It is easy enough for it to be a 'pre-' or post-' job or even to do this inside your current job. I've been doing this a long time and have never had a need to fall back on this ODBC 'technique'. If this wo...

[Gomez] "Tish, you spoke French!" [/Gomez]

My dba told there is no way to increase the size of the rollback segement since they dont have the privilege to do it. What? Talk to whatever 'DBA' you've got that does have the privs. Another option is to run your processes when the tables you are updating aren't quite so busy. A 'snapshot' error ...