DSXchange

That is why I asked if the '+' is in the data. Add a DUMMYPLUS1 and DUMMYPLUS2 (or FILLER) column of length 1 right before the PIC 9(2)V9(6) columns.

No, your two level 05 items are just group names without pictures; that is why I suggested you remove them.

Hello and welcome to DSXchange.

I use the first day of the following month, then subtract a day from that.

DateFromDaysSince(1,StringToDate('2007/12/01','%yyyy/%mm/%dd'))

Functionality notwithstanding, performance can be dramatically faster with the Enterprise stage, particularly with a partitioned database. I would not use ODBC given a choice.

As long as you check for null first and just use OR clauses you can do this in a constraint. As soon as you have an AND then all conditions are evaluated and any normal operation on a NULL field will cause an error.

As far as I know, "24:00:00" is an invalid time, the highest legal value is "23:59:59". If it is coming in as a string, you can check and replace it with 00:00:00 or 23:59:59 depending upon which interpretation you have of that value and then do the conversion.

DataStage hashed files have a configurable maximum length; I believe the default value is 768; you can check this value by looking into the uvconfig file or executing the command 'smat -t' from the command line. This would mean that the hashed file keys are truncated to this length, which explains w...

Lokesh - 'auto' DOES do partitioning. What happened when you made the reference links 'entire'?

The first thing I would do is change the intervals, those are very, very small for the size of data you have (at least 2,800 commits!)
Also, have you checked whether this table has triggers or indices or constraints?

Sorry, I read your first reply and saw the 'yes' to the NFS mount question. If telnet works, then rsh should work as well. With that command you could check for the existance of the file.

Can you get the java into a library routine? If so, then you can create a c++ interlude call and put that into a library and call it directly from the transform stage - this is much more efficient than pushing and popping a shell & executable for each row of data. The record itself can be interp...

Yes, 3 separate user variables in that activity (I'm fairly certain that user variables can be used later in the same stage, analogous to what happens in Transform stage). I would write this as a simple routine CheckRowFunction(DummyArg) JobName = 'MyJob' LinkName = 'MyLink' StageName = 'MyS...

What limit does Netezza impose on column names? DataStage certainly can handle much longer names.

The problem at 7.0 could have been a bug that is resolved at 7.5.2; but the workaround you described should still work at 7.5.2. What exactly is the question that you have?

Don't you think that telling us what the error you received on the COUNT command might help? The HFC.exe is (I think) on the installation CD-ROM. But in your case it might just confuse matters. Wait until one load is finished, then do an "ANALYZE.FILE {YourFileName}" and post the results.