DSXchange

Thanks for that; it's really interesting to have some idea where folks are. Can I ask folks to keep your actual whereabouts up to date? For example, though I am based in Australia, I can now keep people informed when I'm in some other country.

Welcome aboard. :D

Create a bat file that executes dscmdexport (maybe renaming yesterday's DSX file first). Execute the bat file via AT.

Yes you would need 17 Lookup stages. Each lookup is based on the key (read "primary key") and can therefore only return only one row.

Have you looked at the chapter on the Modify stage in the Parallel Job Developer's Guide? Only beware that it has the wrong function for handling null (which won't affect your current issue).

You shouldn't mix JCL with data. You can do this outside of DataStage (perhaps using type or copy command). But DataStage will object to the "rows" of JCL because they don't match the metadata that describe the data you are otherwise processing.

Hashed File stage can only do key-based "=" lookups. If you really do need to lookup against non-key columns in a hashed (not "hash") file, you can use a UV stage. It will be slow unless you index the search columns. Search the forum for more information.

One job, but with multiple Transformer stages (say, not more than four lookups per Transformer stage) would also be OK. You can enable inter-process row buffering (and, if desired, interpolate IPC stages to make it obvious that there are separate processes involved). Of course, splitting into multip...

Seems like an ideal application to use an External Filter stage.

My OS is Windows and is there any alternative for /dev/null? The reason for using TRANSFORMER stage is, my table has timestamp fields and it has NULL value also. So direct ODBC -> SEQUENTIAL gives the fixed length NULL value error. A file called NUL exists in every folder. So you can use .\NUL as t...

I would expect to see "license count exceeded" error message in this case.

Did you change the length attribute of any VarChar (string [max=n]) column? If not, report the bug through your service provider.

There is no S-FTP stage such as you postulate. You could create your own, as a Wrapper stage.

You can use partition parallelism in server jobs. Enrol on Server to Parallel Transition class (offered through this site) to discover how.

Theoretically at least fixed width format is always faster than delimited. The main reason for this is that, in delimited files, the process must read one character at a time, checking whether each is a delimiter character. With fixed-width data, data can be processed (a) in row-sized chunks and (b)...

On the server, at the operating system prompt, type this command. oerr ORA 02291 This will (a) prove to you that it's an Oracle error (not a DataStage error), (b) suggest what you can do about it. Then you can modify your job design so that you do not violate a foreign key constraint - maybe by load...