DSXchange

What the heck is your goal here?

dsjob -jobinfo

Search the forums for "64BIT", that should turn up quite a number of conversations on the subject of large hashed files.

There's no published metadata for the repository tables, but a "LIST DICT tablename" might help. A little.

That does seem odd to me too, but then I don't work with SQL Server so aren't aware of the finer points of its use.

Pretty much everything you'd ever want to know is in DS_JOBS and DS_JOBOBJECTS.

So, only one column actually supplies values to the where clause and the other three values are hard-coded? Then you only need one key column. The number and order of the key columns must match the "?" bind parameters used in the sql.

Don't confuse 'key columns' with indexes or indexed columns. Mark as a 'keys' the columns needed to drive the deletes, i.e. be included in the where clause.

So... you were not in fact logging into the database as the same user?

A job's "log file" is a hashed file and a blink error in a hashed file is a blink error, even if it's not a log. You've corrupted the hashed file by trying to write more than 2.2GB of data to it, you'll need to switch it to 64BIT.

Binaries = executables.

There's no "bulk" option in the DRS stage. What post? Can you include a link to it so we can see the context of the discussion?

Then wrap the ends of the string in your delimiter (pipes here) and then look for "|Test|" rather than just "Test".

Code? There certainly are usage examples of the DSGetLinkInfo function in the online help.

I don't see any mention of a 'compilation error' in the original post. Please start your own topic, post your job design and don't hijack other people's threads.