DSXchange

How many nodes are you running these jobs on, are they using the same configuration file and on which column are you partitioning your data?

Do it step-by-step. Write a job sequence that calls your routine. Once you have that working, put a nested condition in the sequence that branches depending upon your routine's return value.

Could you explain the rule for your example? One can use the TRIM() function to replace double-quotes with single-quotes or to remove leading or trailing or duplicate internal double-quotes, but in your case it would seem that you have different rules.

Is this in one field, i.e. in one string and duplicates are defined by commas?

What database stage are you using? Usually you can choose a write mode of "Delete" so each key that is sent to the database stage deletes that row. So in your case it would be a job reading one database, then splitting into 4 streams (copy stage) and doing 4 discrete database table deletes...

On the face of it, it sounds like you need to write a job which reads from your rejection table and perform "delete" actions on your 4 staging tables.

That would be very basic Job Sequence work. Execute the routine, use the return value to decide whether or not to execute a job from the sequence.

There are 2 problems with your code. DSGetLinkInfo(DSJ.ME,"Flatfile_Stagename","Flatfile_Linkname",DSJ.LINKROWCOUNT) The other problem is that this gets called for each row, so your first row value will be 1, the second row 2, etc. This would be more efficient by using @INROWNUM ...

I seriously doubt that the bottleneck is doing in-memory type conversions. Run your job again and look at CPU usage and I/O usage and you will find that the longer runtimes are because more I/O is being performed. Writing VarChar columns with defined lengths to datasets sees DataStage using the full...

You will need to talk with your UNIX administrator. It seems that either "rsh" is not installed or not in your path, likewise will most likely be the case for "ssh" as well. Once you sort out getting this command working (with or without adding userid/password) then you can imple...

In that case you need to ensure that you don't get deadlocks when running your script in parallel by either committing everything immediately and/or by using different records in each script.

Please explain what isn't working? Is it unable to open the file, or doesn't read a record? Your code really should cover both possible errors, i.e. Ans = "" DFILE = TRIM(Arg1) OPENSEQ DFILE TO F.DFILE THEN READSEQ REC FROM F.DFILE THEN Ans = "Y" ELSE Ans = "N" END ELSE...

If you specify column lengths, then give them realistic values. Using Varchar(1024) for a column that contains "Y" or "N" values will make jobs run slower as DataStage will write the fixed-length maximum of 1024 for each row to your dataset.

I gave you the syntax, you need to add the "default=0" string, e.g.

You need to create a default value for DOLLAR_AMT. In your schema file add "default=0" or something similar.