DSXchange

As mentioned by ArndW, there will usually be a batch ID loaded to all the tables, unique for every run. And in order to revert, there will be a separate delete job, which will pick up the batch IDs and delete all records for that batch ID. We also had a situation when there was no batch ID and the t...

Sorry. I'm not able to understand how your file looks. You use CFF if your file is in EBCDIC format. If you are not sure about the format/structure of your file, ask it to the guys who create it. Understand it. Then I hope it will be easy for you to read it inside datastage.

How long is your text column actually in the file? I remember using longvarchar to read from CLOB or to read the source XML files.

A sequential file stage.

So if the file is ASCII, why are you using CFF?

You did not answer the important question. Is your file ASCII or EBCDIC ?

You have the same metadata from both the links right? Use funnel.

Okay.. Sounds like a kinda dynamic pivot :D If you not gonna be dealing with a lot of rows, write your records to a file as csv. Read the entire record as one single column. Your ground zero year is 2012(You can even hardcode this value) and add one to it for every iteration to get the year and put ...

You mean to say only the value of the years will change and the number of columns will remain the same? What is the indication you will have w.r.t the year? I'm not sure if pivot can handle this, though the first thought comes into the mind is pivot after reading the requirement :lol: As you are in ...

Pic s9(4) is not the same as Pic 9(4) COMP. And there it is.. It should not be a binary field on an ASCII file. Is there a reason why you are trying to read ASCII file using CFF? And not a sequential file?

If you gonna have only two years, why don't you split it in the transformer? Two links out of your transformer, one with first year and the other with another year and then join them again? Am I missing something?

i have defined FORMAT_CD as BINARY [4] USAGE COMP only its in output its as INT 4, rest all columns are CHAR not aware about the NLS map what it should be..? So that column in copy book is mentioned as Pic 9(4) Comp.? And NLS, if your source file is in EBCDIC format, you should set it to the EBCDIC...

Do you have any packed fields? If so, how have you defined them?

And are you trying to read more than one record type out of the same CFF stage? If so, did you try deleting other links and view the data?

What is the NLS map you have defined?

How did you define your metadata in CFF? If you have defined it manually, try importing the actual copybook.

If you don't include the namespace in your X-Path, it will not fetch the data. Maybe one of your files is defined with namespace and the other is not. If this is the case, the same X - Path will not work for both. E.g., With namespace: /ns1:root/ns:1records/ns1:value/text() Without: /root/records/va...