DSXchange

From the looks, it must grab the correct row.

Are there any successful lookups also? If yes, try to find what is common in the successful ones which is not in the failed.

The catch here is - what do you mean by "from production environment" ?

You can split the logic in dd/mm/yyyy or dd/mon/yyyy as below In both cases: yourYear = MatchField(yourField, "0X4N0X",2) In mon: yourMonth = MatchField(upper(yourField), "0X'JAN'0X":Char(253):"0X'FEB'0X":Char(253):.......:"0X'DEC'0X", 2) You need to decide fo...

Are there any ds phantoms or osh running ?

Wonder whether the final result will make any sense !!??

But one possible solution can be

Source to be split into 2 sorts and add rownum in transformer followed by join by rownum.

What will be the volume? If it is a small source file, you may be better off doing in Unix itself.

If you have 10 rows as
1 1
2 2
...
10 10

do you want it as
1 10
2 9
3 8
....
10 1

Is this assumption correct ?

What if you remove constraint from second link and enable 'otherwise' ?

Please define duplicates.

Are you looking for identical rows or duplicates by specific columns ?

search for isNull or NullToZero

Simpler method may be to do in sequencer.

You can read the list of files in execute command stage followed by encrypting the date in user activity and conditional stage with custom logic for dates.

If you want in Unix, you have to do your own date maths for date difference.

Will you be communicating across the two machines at any point in time? If not, you can try doing it (other than datastage license) xmeta connections will be the main one to change. Also you might have to look into the ip address of the two machines if you are going to connect from same client. Not ...

Add a copy stage and do not propogate.

amitendra_amit wrote:All of sudden all the datajobs in our production environment has started running slow.

Sudden ??

Wonder what is the purpose of quoting those strings ?

Errors ? Logs ?