DSXchange

Ack... parallel job. Yes, what Arnd suggested could be an option as well, depending on how tight that 'process one by one' requirement has you in its grasp.

Use a combination of DCount() and Field(), the former to let you know how many 'fields' there are and the latter to pull out everything from the 2nd field to the 'next to last' one.

In other words, a looping Sequence job.

There's not much to ask. You cannot process XML files that large and people that send them out like that should be shot. IMHO. :wink: You'll need a 'pre-processing' step to break them up into digestable pieces or use a third-party parsing tool that doesn't need to suck the whole thing up into memory...

'Read' in what sense? If you are required to process them one by one, build a looping Sequence job. Otherwise, cat them all together (before job or in the Filter option) and 'read' them as one big file.

By default, Sunday is considered the first day of the week and there are places where you can pass an optional 'origin day' to change that. This function doesn't look like one of those. :( Seems to me you'd have to run some tests to determine how it decides what a 'week' is or you'd need to ask your...

Is it working for you? If so, it is certainly *a* right way.

asorrell wrote:Craig? Are you getting too old to chase tail?

My own tail, yes.

Don't have to be a Moderator to do that, Andy. Well, unless it's much later and there's a reply, then you would need superpowers.

OK, so finally a much more representative example gets posted. Let's keep going... What needs to happen to this record? I get the 'ignore/remove the HDR and TRL information' (right?) but there are 9 DTL pieces stuffed in there. Where do they go? Do you need to parse out them and summarize them in so...

Can I humbly suggest that we wait for the requirements to be clarified before we throw out any more suggestions? Especially UNIX based one when this may be easily accomplished 'in tool' once we know. Thanks.

U is not here and please do not advertise your problems via private email.

No-one can help you without details about your job design and data. What stages / database / settings are you using? How are you assuring updates happen when they should? Details, man, details.

Sreenivasulu wrote:As far as i know AIX uses JFS file system and it does not have a limit on the number of files.

Ray said the same thing: "no practical limit on the number of files". The issue is specifically the number of subdirectories supported.

The Index() suggestion is akin to '%DELL%' rather than 'DELL%'. If the condition is index()=1 then it will work like 'DELL%' and index()<>0 is similar to '%DELL%'. :wink: True, but my point was that it does a substring match across the entire string so can find it anywhere and without you specifica...

You can't do a "load" as that uses sqlldr and bulk loads don't commit in the sense that you need. And you need to esure the job aborts if there are any rejects so no closing commit is done.