DSXchange

We can not provide you "syntax of restartability". This must begin with a specification. You can certainly code branches (though there are those who would frown on your use of GOTO since such a jump is itself irrecoverable). But when, where and why must be determined by careful analysis of what you ...

You may need to convert to String before applying a Trim() function.

It still exists. IBM will push you to purchase version 8.0 (IBM Information Server with DataStage), but the hardware requirements are substantially larger. IBM will offer to sell you more hardware, which is the only solution.

How have you "come to know" this? It's actually true, and its purpose is to make data transfers occur on whole-word boundaries. This, in turn, means that the data transfers can use more efficent low-level functions such as strmov(). But that's all totally transparent. Don't worry about it. The Hashe...

If the physical size is still large after an auto-purge has occurred most of the internal space is unused and will not affect the fact that the effective modulus is small. If you really want to reduce the size, then RESIZE RT_LOGnnn * * * USING /tmp is the best command. Use any file system with free...

You don't need to change the code at all. Just disconnect between tests.

Variables in COMMON are persistent. The routine has been written to make use of this fact, to persist the array of match pairs between rows.

What your problem is is merely an artifact of your test strategy.

Ascential DataStage version 7.5.1A

Yes

What output did you get from DS.CHECKER? (The output is preserved in the subdirectory &COMO& in your project directory.)

My t-shirt is somewhere in the baggage retrieval system they've got at Heathrow

Stop being so certain.

The some thing might be an automated "cleanup" script or a disk utility gone slightly haywire.

My money would still be on someone rather than something.

Getting them to admit it is a whole different game.

Try PRECISION 14 (the maximum legal value).

Ah.

Don't have that.

Welcome aboard. :D

The short answer is yes; with no purging the logs grow indefinitely (until they break at 2GB) and this increases exection time.

Applying new log purge settings will have no effect until the job next completes successfully, because it is this that triggers an auto-purge.

Copy stage would be the most efficient to effect the column name change.

The reason for your "problem" is that the collection of column definitions is a property of the link, not of either stage.

Obfuscation? Maybe there are lots of #define statements literally changing the syntax to make it look like BASIC? You can do the same in BASIC with EQU LIT.