DSXchange

Replace existing rows completely should, for each row processed, delete THAT ROW then insert its replacement. If you don't have a key column defined, so that no WHERE clause is generated, each such delete will delete the entire contents of the table. So check that you have at least one key column d...

Check that you have UPDATE privilege on the table in question. Your Informix DBA will be able to verify. Check also that you are identifying at least one Key column, and not supplying NULL in that column. Inspecting the generated SQL will help here; if there is a key column, then there will be a WHE...

It's easy enough in DataStage - if the target table has the extra column, simply use the new metadata (table definition) an assign the constant 1000 to that column for all rows. Probably easiest is a Column Generator stage with a cycle beginning at 1000 with an increment of 0.

Welcome aboard. :D

The short answer is YES.

Techniques can be found by searching the forum (perhaps use Type 2 and SCD (slowly changing dimension) as search terms).

I think this is unrelated to the earlier problem and possibly deserves a separate thread. The message Variable 0.9 could not be found in data dictionary suggests that you have - whether advertently or not - created a field name of "0.9". Because it is from B this suggests that it is in your second, ...

Sometimes it is necessary to restart Windows (where have we heard that before?!) for environment variable changes to take effect.

Explicitly double quotes. This turns them into "quoted identifiers". You also need quoted identifiers if they are crazy enough to require illegal characters (such as space) or case sensitive identifier names.

"keyword YEAR" suggests that YEAR might be a reserved word. Apart from that the load script looks OK to me; you might edit the post and disable smilies so that 8) is not converted into the "cool" smiley.

There's no mechanism for this in job sequences, unless you can come up with a regular expression or command (script?) that turns the file into a delimited string. You could, as an alternative, build a server job that invokes UtilityRunJob routine for each row processed.

Find the "active stage finishing" entry in the log - this contains all the link row counts. Otherwise use DSGetLinkInfo() function in the DataStage API in your own routine.

Welcome aboard. :D
1. Save one or more commands from the history stack in the Administrator client's Command window.
2. DS_AUDIT for date/time created, RT_LOGnnn for the log. Search the forum for more information - it is there.
3. Yes. Again, it's definitely been posted in the past.

OK I understand now. Is it exactly three flat files? If so, the easiest way is a Transformer stage with three outputs, and constraint expressions based on the first field. You can use the Field() function to extract it from its row if needed. You can run this job then the following three from a job ...

Use stage tracing to capture some of the data into a trace file.

But complain, nevertheless.

That's entirely in the hands of the folks who run this site - they own the class and run it on a fairly regular basis. I believe the next one is in the first week of July - watch the forum for an announcement.

Look at the sqlldr control script, particularly the line immediately ahead of YEAR. Is it complete? Or is there a number missing from the end of it?