perfomace of lookup

Raghava · Post by **Raghava** » Thu Apr 06, 2006 1:23 am

hi everybody,

for lookup which one is better hashfile or ODBC stage or Oraoci stage? and why?
do we have any unconnected lookup in data stage?
how to handle unconnected lookup logic?

thank you
-----------
Raghava

kumar_s · Post by **kumar_s** » Thu Apr 06, 2006 2:09 am

Depends on you desing and requirement the stages can be decided.
Datastage doesnt have any unconnected lookup in data stage like Informatica have.
But you can relialise similar performance by populating the hash file and it can be lookep up at multiple places.

chulett · Post by **chulett** » Thu Apr 06, 2006 6:55 am

Seeing as how hashed files are specifically designed for use as reference lookups, the answer is pretty much always hashed files. There can be specific instances where a database stage can be 'better' but under extremely limited circumstances.

What exactly is the definition of an 'unconnected lookup'?

kaps · Post by **kaps** » Thu Apr 06, 2006 9:57 am

Unconnected Lookup is a Tranform in Informatica. I would say it's similar to Routine in DataStage but not exactly the same!

We can have that stage(unconnected lookup) in our mapping or job which will have input columns and output column. We can call this unconnected lookup N times from that job just passing input column values to it and it will return the column which is designated as output column. This caches the table or file in memory when the first call is made which improves the performance.

Connected Lookup is something you need to have N times in the map if you are calling it N times.