DSXchange

It's also possible that the environment variables for Oracle (particularly ORACLE_HOME) are not correctly set for the process executing the DataStage job, so that it can not find the tnsnames.ora file.

In a server job, Oconv(Iconv(InLink.TheDate, "DYJ"), "D/YMD[4,2,2]") In a parallel job you could use this expression in a BASIC Transformer stage, if you didn't mind the performance penalty. Otherwise you will need to use parallel job functions. Why not start a ne...

Depends what you want out of it. Why not start a new thread on the parallel forum?

Assuming you can read the Binary(8) column at all into DataStage, you may be able to convert it using Iconv(InLink.TheColumn, "MB"). Try reading it as a Char(8) if Binary(8) does not work, in which case use "MX0C" as the second argument of the Iconv() function. That will give you the content. You wi...

SQLSTATE is specific to ODBC. It's the "generic ODBC error code". It is not populated by any stage other than ODBC, UV (which uses ODBC protocol) and Dynamic RDBMS (using ODBC protocol).

You can supply these to the job doing the updating as job parameters, or calculate them within the job. The actual format that you need to supply to the database will depend upon which database you're using, which information you opted not to reveal. But it should be a straightforward update execute...

Then maybe your problem is in the other ODBC stage (or DRS using ODBC protocol)?

Talk with your DBA. It is DB2 that is unable to allocate enough storage in the query heap, not DataStage. The suggestions others have offered are likely to help, but you may not have sufficient privilege to execute them. You also need to make sure that they are executed in the same process as the qu...

It is also possible to tune the memory allocation of the Aggregator stage on a per-job basis, but sorted input renders this unnecessary because memory can be freed as soon as any of the sorted values changes.

It is created in the Data folder, along with all the other data files.

You can also ask for it to be retrieved to the client, but I find this cumbersome. Editors like UltraEdit can FTP the file from the server more readily.

Thanks for posting the resolution. Now how about marking the topic resolved?

In a server job, simply pass it through a Transformer stage in which the input link carries the column as a VarChar and the output link carries the column as the appropriate numeric type. Ideally, constrain the output link with a test that the value is actually numeric. The Num() function is ideal f...

Why me? I keep telling you to use the job's user status area, but you insist on demanding an answer that uses a file. The answer is: read the file. Use an Execute Command activity, capture its output ($CommandOutput activity variable) and take the first field of that dynamic array as your count.

What is your "triggering tool"? There is usually some startup overhead, more in parallel jobs than in server jobs. There is also a delay before the Director client refreshes. What evidence do you have that job A is "waiting"? How do you measure this? Is ANY of the processes in job A (for example DSD...

Not using metasyntactic variables (placeholders in syntax representations) instead of genuine operands in expressions will help too.