DSXchange

Thanks loveojha2. For some reason i thought it was to be incremented for the same group , but if it needs to remain the same then yes, you fixed my code. Just a small change is required in this one. The first row needs to be handled as it has nothing to be compared against and the stage variable 'co...

Thats true. The OP can just check for change in the col Z. It should be ok as long as its sorted properly. I just translated the OP's code in which he is checking for all the three columns.

Welcome Aboard,
You can use either the ODBC stage or the DRS stage. I am not exactly sure if PX is equipped with DRS or not, if it is, go for that.
Regards,

You guys are just awesome. Ray's name speaks for itself and ArndW, cant find words to describe your importance to this forum. Congratulations once again, on behalf of dsxchange and me , on blazing to the post numbers you guys are at today. Honestly, we all lookup to you guys. Thanks for teaching us ...

One problem of doing this with a transformer is that the file's name has a date and changes and looks like you can't use wildcards on sequential files Thats a problem, yes, but whats stopping you from creating a second file with a constant name with just the first two records (constraint of that li...

grep -c will only return 1 if there are no line terminators. The OP has still not confirmed whether is a single huge line or multiple lines. For that he needs to perform a sed operation to replace the martian delimiter with a LF and then simply do a wc -l. That should do it. Something like sed 's/@@...

We have some TX gurus here as well. Keep up the good work janhess

Probably create a small script that sources the dsenv file and then executes your job.

Try this, counter: if (TRIM(X):TRIM(Y):TRIM(Z)) = prevVal then counter + 1 else 1 prevVal: TRIM(X):TRIM(Y):TRIM(Z) Use the counter as the derivation of nO. Also make sure the incoming data is sorted on X, Y and Z, in that order, ascending.

A piece of advice. These questions are more appropriate for a Unix forum rather than DataStage. I am not saying we wont help you, dont get me wrong, what i am trying to imply here is that you will get answers more quickly or even find them by googling it much faster.
Regards,

You can build a simple DS job to create a sequential file with space delimited email addresses. Use that file in your unix script. Basically a small script that fires of the DS job, reads the file and sends email.

Ok, try this,


That arithmatic manipulation should get the issue resolved, then again i do not understand why you/data modeler, has set an amont field as varchar in the database.

What database are you using?
If its in oracle then you can use the Lag() and Lead() functions to get the previous and next values respectively. Look at the following website on more details on these functions
http://www.adp-gmbh.ch/ora/sql/analytical/lag.html

In such cases, i would love to build a hashed file and run sql's on it via the universe stage. For small amounts of data, its pretty smooth.

I do believe you'll find that DB2 wants any date fields in internal format... I wonder if the same is true for timestamps? For timestamps it takes in regular format (not internal). Internal is specific to date types. As Ken suggests, try loading it with sql type varchar. Also note, if you are tryin...