DSXchange

Which is why the original answer posted 2 months ago used the mnemonics. Never mind the fact they are much more obvious as to their intent than just the numeric values.

Then write and use your 10-line perl script. No-one is forcing you to use the Merge stage, I for one find it unwieldy and have never used it in the umpty years I've been working with the product. And not sure why one would expect an ETL tool to work like a BI tool. [shrug] Did you supply a temporary...

You are out of luck if they are pathed hashed files, but if you created VOC entries for them... in essence as if they were in an 'account'...

eostic wrote:Verne and Ernest (some off-beat country bumkin --- comedy or a series, not sure which)

The late Jim Varney played a character named "Ernest" and his registered trademark catch phrase was "KnoWhutImean, Vern?". Yup, spelt just like that.

The limit in DataStage is how long it will wait for something to start before it gives up.

It creates two hashed files under the covers to hold the information from the files and then performs the join between the two data sets as you've configured. Sounds like you may be crossing the hashed file's 2GB barrier here with your volumes. Is that the kind of "merging" you really need to be doi...

Search the forums for "century pivot", I found at least two posts with a possible solution in them for you.

Look again: Manager/Transforms/Built-In/String. If that won't work, you'd need to provide us with your spec/requirements stating how you will recognize any given date in your text. Then we can help you implement those rules.

I'm gonna have to check out the "historical Ernie" list, it seems.

If you just want the numeric portions, then the DIGITS transform will do that for you:

DIGITS(claims.11122007.xml) = 11122007
DIGITS(claims.2007.11.14.15.20.32.xml) = 2007114152032

Etc. Then play from there.

np. The second URL came from the first and are the specific requirements for Windows. All platform requirements are available individually from the first URL.

Looks like this is a good starting place:

http://www-306.ibm.com/software/data/in ... ments.html

And it also looks like they meant to send you here:

http://www-1.ibm.com/support/docview.ws ... 23&rs=2321

Ask IBM to send you a working URL, then share it with us.

Suggest you tell us what you need to do with this file - what bits do you need from it and what will you be doing with those bits.

Look, you lucked out - Uncle Ernie back from Holiday Camp so soon.