DSXchange

If it is specified as a date field, dont you have to use StringToDate() on your string?

bcarlson wrote:... and you'll watch your hair growing faster than your job is running. Okay, slight exaggeration...

No. For multiple files, multiple nodes(if available) are used. All this is documented in Parallel Developer's Guide.

Ok now thats what I was doing wrong. Bummer. Thanks Craig.

It worked for list huh? Boy, I have to try that again at my end.

What version of Informix are you using? I heard that for most versions, the size limit of column names is 18. Did not confirm it though, you can google it up.

Craig, are you using a counter or a list? I tried with list but for some reason its not accepting any functions like Field() or Convert() etc; in the list specification.

Well, lets try to confirm, that really is the issue here. Split the columns into half and see if the job compiles. As Ray said, it could be a missing quote, which is something totally different. So lets try to cross out as many items as we can here.

I believe there is a new line character at the end of the list. I tried doing this for a fellow collegue but removing the @FM character at the end does not work while specifying it in the list. You will have to preprocess the list, remove the @FM character and send it as a job parameter to be suppli...

Well, that might be the problem. Test it out. Cut it into half and see if the job compiles. If to no avail, cut it into half again.

Well, my advise would be to pre-process the file, either with datastage or outside datastage (awk, sed, shell script etc) and disect the file into individual files. Once the files are split, based upon the table, it will be easier to read/load them.

As I indicated, search will reveal all, as all this has been discussed time and again. Search on keywords like "64 bit hashed file".

Search the forum for examples of vertical pivot. It can easily be done by the classic approach of doing key comparisons between present and previous key and storing the values in stage variables.

Can't you regenerate it?

Yes. I read that before. My point is, go to the database (not from datastage) and verify if all the columns do exist.