DSXchange

ray.wurlod wrote:I understand that DataStage regards 00:00:00 as if it were 24:00:00. So you'll need to test for the date. ...

That works. I check the time part and when it is 00:00:00.000000 then i subtract 86400.000001

I have a requirement to subtract .000001 microsecond from my current timestamp. I have tried both the TimestampOffsetBySeconds and TimestampOffsetBySeconds functions: TimestampOffsetByComponents(StringToTimestamp(DSLink4.timestampraw,"%yyyy-%mm-%dd %hh:%nn:%ss.6") ,0,0,0,0,0,-0.000001) Tim...

No,

I've moved to a new client. The last thing we did was change the auto purge setting. The problem might of went away.

Sorry!

Good catch. I should read the entire post.

You will find it in the transform. Stage properties for the reference link.

Do you have the "Reference link with multiple row result set" box checked?

I like to use an after-job subroutine and the DSGetLinkInfo. I'm sure others will suggest to use UNIX utilities.

Sorry to not get back to you sooner. The tech supporting my case just keeps sending me more questions. I will update the thread if I ever get an answer.

I would recommend that you read the Parallel Developer's Guide.

I have the exact same case open with IBM.

Could be contention in you own process. Tell us a little about the design.

Are you using RCP?

Do you know what the partition and join keys are when the datasets were written? You can partion all Datasets by Seq from the start. This will avoid to repatition before the second join. You can still sort the data on Seq and Date.

Try to use the stagename before the container or use a link in the container with the container name first.

Ans = DSGetLinkInfo (handleJob, ContainerName.StageName, LinkName, DSJ.LINKROWCOUNT)

Somehow, the key has to allow for two records to be in the database at the same time. What makes the rows unique in the database?