DSXchange

This is not even a multiple row lookup. It is custom SQL only.

Sorry, I misunderstood.

Parameters do not work in Command Stage except in version 7.5.1. You need to switch to Routine Activity and run ExecDOS or DSExecute. Try it and let us know. I have only done this on UNIX. I assume it works in DOS.

Sorry, brain needs tea. open "whatever" to FilePointer else print "whatever" You have a syuntax error. ID or some key value needs to be sent to this routine unless I do not understand your problem. I am assuming you have some key which is assigned mutiple PRODUCT_ACCOUNT_NUMBERs....

Sainath I believe he is correct as we need to make it adaptable to UV stage to facilitate this process. This is correct but what most users do not understand that a hash file stage can build a hash file which is used in a UV stage. So the hash file itself can do a range lookup. The hash file stage c...

Agreed but as Px gets more transformation capabilities then server gets more speed and multiple threading. I doubt if Px can match server in transformation but neither will server match Px speed. Maybe hardware speeds will make Px less and less neccessary. The flexibility in server is worth the diff...

I think Sainath is correct in saying we need to see your SQL.

Each SELECT may need to update a hash file. The number of rows in all SQL statements would be nice too.

Px is better at E and L in ETL. It is very fast if used properly.

Server is better at T in ETL.

Now you know who put the T in ETL. Px requires mutiple processors otherwise it is a waste of money. In my opinion.

There are lots of options. Hash files are easy to do outer joins with. Depends on lots of factors on which option you choose. If your source systems are overloaded then you want to do it in hash files or off line as much as possible. Hash files may out perform any other method. It may not as well. I...

Wrong. There are many posts explaining how to do this. You need to make sure your keys do not remove the duplicates. You will also need to use a UV stage and not a hash stage.

You cannot move the DataStage files around. The complete path is stored internally in some critical places. You will need to reinstall. If the path was E:\Ascential and still is then you have some other issue. Has your path changed?

Kim, no Kevin.

I would not process this with ODBC in a routine. MyLookup(ID, PRODUCT_ACCOUNT_NUMBER) Ans = 0 open "HOLDER_UV" as HashFile else stop read Rec from HashFile, ID then NoAcctNos = dcount(Rec<1>, @VM) for i=1 to NoAcctNos if Rec<1> = PRODUCT_ACCOUNT_NUMBER then Ans += 1 end next i end ...

Here are some more good questions: Colin ETL_JOB_HIST,TIME_DIM,ETL_QA_SQL and ETL_ROW_HIST. Q) How would I go about populating these as Hash Tables ? You should have the DDL to create these as Oracle tables. ETL_JOB_HIST should be populated by DSJobReportDb and so should ETL_ROW_HIST. You do not nee...

Ken Bland posted a routine to update all jobs with purge settings. Do a search. Works great.