DSXchange

Its not good. Say you have a stream of 50 jobs. All dependent upon each other. The 49th job creates a staging file thats needed by the 50th job. The 50th job fails for some reason on friday. If you are not able to fix the problem in the next 72 hours, the file will be gone. You need to give enough t...

There is a function present in px jobs, not in ds jobs. The work around is to get it from the unix level by echoing it.

The CDC stage gives an extra column with the codes. You must be getting them, so whats the problem here. Just insert them. Or you want wordings instead of codes? If thats the case then stick in a transformer and change the codes to words.

Thats odd. If you have it defined as nullable all accross and it still rejects it then that means something is missing. Double, triple check in all the input/output tabs of all the stages.

Did you try searching?

No. I doubt there is any other way. Basically you will have to convert to string and then do your substrings which you already are.

Do an exact search on "create report".
The following is a fairly recent discussion.
http://www.dsxchange.com/viewtopic.php? ... ate+report

Out of curiosity, why do you want to change the decimal field to char. Is it just becuase of the leading space. Let it be there, the space will not affect the fixed width file format.

Does your table accept NULL for the timestamp field? If you get rid of the database stage and load it to a flat file with the timestamp field set to null, can you load the data into a file?

If you want to write to a file as Decimal it will always save a space for the sign. If you dont want that, change it to string and load it to a varchar field.

Welcome Aboard
Change it to integer within the jobs and see if the error persists.

You have to take into consideration partition number and number of partitions when dealing with @INROWNUM in px jobs. Refer to this post for a better understanding.

Exactly, and therefore delete is not logically the same as insert. Hence, inside datastage, get a select count(*) and load it somewhere, UserStatus, hashed file, text file; your choice. Pass the delete and do count(*) from the same table again. Do the comparison.

loubi wrote:
I must make a count(*) before and after job execution to know how many lines were deleted.

You have the solution then whats the problem. Get the count before and after delete. Subtract them and thats the number of deletes.

madhukar wrote:
i am developing one more sequence which it checks the status of seq1 and runs job A.

But why? If the next job A is to be run whether previous jobs are successful or failure, then why check. Just run job A unconditionally.