DSXchange

First, rerun the job with APT_DISABLE_COMBINATION set to True and then let us know the exact operator/stage producing the error.

What stage are you receiving this error from? Cannot remember what was available at 7.x, but I think DB2 partitioning was not available as a partitioning method unless it was the DB2 EE stage.

Change your "same" partitioning to partition by the key you are using. Also, change the setting for "Log Statistics" to True so you can see in the director what the stage sees as inserts or edits. Could be that the files are truly not the same. From Unix do an md5sum on both just...

It is not uncommon to have more nodes defined than CPU's.

BOM or Byte Order Mark will be added by default when using Notepad and writing UTF-8. No such behavior is present with Unix since files in Unix can be multi purpose and the BOM may interfere with processing correctness. The BOM is usually at the beginning of the stream to let the receiving appliatio...

You are welcome

A copy with no output is not writing anything

Correct - have no output out of the copy and you will have no issues.

Lookup requires a stream output, but does not require a reject link. If you only have a reject link and no stream output the compiler will complain. Send the lookup success rows to a copy with no output or a seq file stage and write to /dev/null or some such thing.

I am not a huge fan of folks going after the UV repository in such a manner, but you can retrieve a list of invoation id's and that is what you want.

I might suggest that whomever set the resolution status of this thread to "WORKAROUND" change it to "RESOLVED" since workaraound implies that there is a problem with small int and integers in that they cannot be zero padded. Converting it to a character and padding that STRING wi...

thanks where in the sql do we put this? NODENUMBER(partition_key)=CURRENT NODE Also want to mention we have a dpf environment, didn't know if that makes a difference. SELECT COL1, COL2 FROM TABLE WHERE NODENUMBER(COL1)=CURRENT_NODE Choose Uncommitted Read in the stage and yes this is syntax specifi...

in your sql statement (if user defined) do you have the NODENUMBER(partition_key)=CURRENT NODE dbpartitionnum will work too The stage is running parallel in the sense of ETL - one db2 connector pid for each node in the configuration. The sql statement gets handed off to an instane on the dpf server ...

That is in a header so I would think you might wrap it with a C/C++ program? but then you would have to understand the API and create an object reference and a bunch of other things - is there some reason you need to know this outside of looking in the administrator client?

Hi folks, looks like everyone is getting on board with the unbounded varchar idea for reducing memory and disk consumption. Thanks, Doug I would not go so far as to say this. That is certainly not IBM's take on character string handling. The framework of DataStage moves fixed width data regardless ...

If nothing else this thread is good for a laugh