DSXchange

I know, you know... the comment was for the OP. And it's very clearly laid out in the Help available, hence the suggestion they actually read it.

Read the Help. There are only two columns allowed in the Folder stage: 1) The filename found in the folder and 2) the contents of the file.

If all you want is the filename, don't define the second column.

Umbix62 wrote:May someone explain to me why in the terminal hashed file I have, in this field, the value "Y.." (59 00 00 ine hex format) for all records.

What 'terminal hashed file'?

As noted, sounds like a bug that should be reported.

No, not all that suprising as DataStage is a 32bit app.

When i tried to call DSLogFatal in the derivation of the transformer stage i got an error like 'Array DSLogFatal not dimensioned' which is same as one would get if the header files are missing when you call it from a routine. You're right... I didn't mean you could use it directly in the derivation...

Post some examples of these 'warnings' that you are having a problem with. A Reject link should handle anything that has an issue with the database.

Ok, more things to consider to help with the sorting. The Sort stage is good for small volumes but is pretty darn slow for anything remotely considered 'large'. Let your operating system do the sort for you, and leverage it via the Sequential file stage you already have in the job. Remove the Sort s...

I saved that question for you, Ray.

The 'Show instances' option is only applicable to PX jobs, perhaps you are looking at a Server job.

I thought if someone had a set of sqls I could run them in Administrator Command, then I could do a thorough cleanout. Dangerous and shouldn't be the initial approach to take. All you should have to do for a 'thorough cleanout' of the job is delete it. If something very unusual happens with that an...

No, still not really enough... but we'll keep digging. And this: (CRC logic is been implemented to avoid duplicate values) This concerns me, as CRC32 is not a mechanism to 'avoid duplicates' but for CDD or Change Data Detection. Be that as it may... Which part do you think has slowed down? Are your ...

Ok... at least now we have a clue what this is about.

However, that's not nearly enough information for anyone to help you. Post as many details as you can: job design, data volumes, source and target, everything you can think of. Perhaps then someone can help.

It's more of a 'could' than a 'should'. If you need to do this for five columns (rather than the single column mentioned before) I'd suggest you write a routine to do this. There you can easily do X EReplace functions in a row or inlined or use what Arnd provided or anything else you felt was approp...

Perhaps... if something actually changed. However, it's simple enough for someone to copy a job, replace all the metadata with a new table's columns but forget to actually change the name of the table in the stage. First time you run it... Invalid Identifier.