DSXchange

Got an odd problem with this I don't see yet. Just got bit by the fact that adding one second to a timestamp seems to have a 'midnight' issue: 2009-12-08 23:59:59 + 1 second = 2009-12-08 00:00:00 2009-12-08 23:59:59 + 2 seconds = 2009-12-09 00:00:01 :? Edit: Got it! This line: Case argSeconds > 0 An...

Truncate the table and reinitialize the state file, I suppose.

The docs seem to imply that this can still be accomplised in a parallel manner as long as the source data is read sequentially and the paritions are 'perfectly' balanced.

My second post explains how to do that.

I was going to say the manual explains all this but I don't see anything like that in the 8.x documentation, however there is much more information on how the stage behaves in the 7.x docs. No clue why they would have left that out. :? Here's a snippet from the 7.x manual, it comes with some nice pi...

It *is* generating unique keys. Having 'gaps' in the sequence is unrelated to uniqueness and there's no rule that says surrogate keys (sequences) cannot have gaps in them - especially when doing parallel processing on multiple nodes. Most people don't care, will your application break if they're not...

Interesting. It all looks good so far... what is the datatype of the target field in DB2? Unfortunately, I have no DB2 here to validate anything against so if this is specific to DB2 I'm not going to be much help here.

This advice comes from a Linux question but it applies to all UNIX platforms:

http://www.unix.com/ubuntu/114419-ubunt ... erver.html

Post some examples of your actual data. What is the datatype of the target for the StringToDate function? Does it help if you append a zero time to the DateFld before you convert it?

Somewhere in your file you've got at least one value that doesn't match that format. It's also a little confusing that a time portion is mentioned in the error when you are trying to convert to a date.

Still it doesn't seem to like a year that it found.

Charlie, make sure you understand that connectivity from your laptop has no bearing on your ability to access anything from the DataStage server on the UNIX box. Well, except for knowing you've got the userid and password correct. :wink: Being able to use the 'Plug-in Metadata' stages will be depend...

You have to point to the "descriptor" and it's in your error message: /usr/local/dsdata/whs_dss_datamask/data/whs_dss_dmask_2.0/lookup/patient_id_lkup.ds If you had the 'proper rights' you wouldn't be seeing this error. If you meant you can't find the actual datasets and you're certain the...

"Unable to open descriptor file to update: Permission denied"

Make sure you have proper permissions to the .ds file.

Seems to me that a "Block write failure" is either because the disk is full or you have a media error / bad block / hardware issue. You monitored the space while the jobs ran and the error was generated?

Also, is your O/S 32bit or 64bit?

You'd have to ask your support provider about a patch. How about answering my first question?