I have developed a parallel job in which the source dataset has three timestamp columns. The extended property for these columns are disabled in transformer stage to truncate microseconds into the target. The vierd thing is that one of the timestamp column is populating data with microseconds in the target where as the other two do not.
Now we have enabled the extended property for this timestamp column in transformer and could see that data is not populated with microseconds in target.
Can someone pls look into this and explain this very strange behaviour!!
You need to dump the score into the job log to determine why one field is behaving one way and the others differently. To do this add the environment variable APT_DUMP_SCORE as a job parameter and, when the job runs, set its value to True.
It may also be beneficial to inspect the C++ source code that was generated when the Transformer stage was compiled. This is stored in the RT_SCnnnn directory in your project, where nnnn is the job number from the DS_JOBS table.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
@SURA:
I have already tried writing the data into sequential file. There i could see timestamp columns without microseconds, but the problem is when we write it into a dataset.
One important thing is that after transformer stage we have used many lookups where this timestamp column is a direct mapping from source to target (dataset), but we could only see this behaviour in only one timestamp column.
I don't know why / how you are getting value and that too for only one column and not to others!
Even though you have lookup in between, you can enable / disable micro second option in lookup stage too.
If you need micro sec, ensure that you enabled in all the stages from source till target. If you don't want, ensure that no where it is enabled and thats what i can offer.
The Three timestamp columns are CTCSTRGMTS, BUSINESSPROCESSGMTS, BUSINESSPROCESSENDGMTS. length is 26 and scale is 6 in source. Later the lenght and scale are taken blank in transformer and the extended properties are unchecked. Only BUSINESSPROCESSGMTS column is being populated with microseconds in the target whereas the other two dont.
