DSXchange

Hi,

I'm new to this forum, I found this forum very beneficial and have got some good ideas from posted questions and answers.

I would like to know if anyone knows that is there any disadvantage to use Basic Transformer in Parallel job which is available in the version 7 along with Parallel transformer. I mean is there any performance issue?

My 2nd question is there any PX routine which work like server HashLookup? Since PX doesn't have hash files, it has DataSets, what's the best way to use DataSets like Hash Files and have common routine to lookup these Datasets

Please advise....

Thanks,

Sohail.

There is a small performance overhead in using any additional software. Therefore it follows that there will be a small overhead in loading the BASIC engine. That's where the main overhead will be - actually loading the engine. A small extra overhead in process management, but PX is doing lots of that anyway, you won't notice it.

The nearest PX equivalent to server's hashed file lookups is the Lookup stage, which uses a memory-mapped dataset to perform lookups. Sound vaguely familiar? (Suspect it doesn't use a hashing mechanism to determine the location in memory; but pretty much any memory access is quick. Maybe someone with Orchestrate knowledge can contribute info on how a "get row by key" is managed in a dataset.)

Thanks Ray, I wanted to have second thought before making any decision to use Basic transformer in PX, initially it seems very attractive to use this as it has many built-in transformation routines which are not available in PX transformer yet. But now if there is some overhead involve with it then it may not be a wise decision to use it.

Here is the response I got from Ascential support

"I checked with Eng, and they confirmed that there's no real overhead in
using a BASIC transforemr, although it is highly recommended that you use a PX transformer if you can, and BASIC is normally used when you need to run some BASIC routines."

For all practical considerations, the overhead is negligible compared to the amount of work that PX will be doing actually processing your data. Don't feel constrained by the tiny overhead involved. Contrast the amount of learning and work that you'd have to do to implement the same rules using the PX Transformer stage!