DSXchange

chulett wrote:Use two parameters, one for the folder and another for the filename. ...

Is there no other way using one parameter? Making two parameters is currently the last option as it will involve many other changes for us.

Is it possible to manipulate the a job parameter in the Folder stage? For e.g. I have a job parameter "File" which has the absolute path to a file e.g. File = /a/b/file1.dat Now in the folder stage in the Properties tab I need to specify just the folder "/a/b" I tried doing #File...

It ought to be possible, but you would need to guarantee that only one file matches the regular expression if you are planning to use it as the filename property directly. Otherwise, specify a filt ... Yes only one file will match the pattern. If so is there any way to achive this? In case I use a ...

I have a datastage job that reads a file using a Sequential File stage and processes it. Now the file name is dynamic and follows a pattern: feedata_<month>_1_<yyyy>.txt e.g. feedata_January_1_2007.txt Does datastage provide any option so that I can provide feedata_<month>_1_<yyyy>.txt as job parame...

If my design is as follows: [Transformer] > [DRS] And all am doing is issuing a delete query from the DRS stage(using the user defined sql tab, and database is oracle): "delete from A where date1 < somedate" Will setting the Array Size and Transaction size help from a faster query executio...

One target stage with before sql DELETE FROM TABLE B where B.KEY IN ( SELECT A.KEY from TABLE A) update option set to "Insert rows without clearing". Source will be your TABLE A. But that is going to take time for 2M records. You need to go for TRUNCATE and BULK loads. I thought of this. ...

Scenario: Two tables A and B, both have columns col1. Both tables contain approx 2 million records. I need to delete all records from table B which match with table A based on col1. Then insert all records from A into B Whats the fastest way to achieve this in DataStage server job? Database is Oracl...

So it is a known fact that if a job uses IPC stage we cannot do rollback if the job aborts? Why so? (I assume this happens because IPC creates two processes?)

narasimha wrote:What is your target database?

Oracle

DSguru2B wrote:Get rid of the IPC stage.

Is there any other option? I need the IPC stage as I can see faster loads with IPC in place.

My job looks similar to as follows: [IPC] -> [DRS] | | [SequentialFile] --> [Transformer] | | [IPC] -> [DRS] Note in the job, the ipc links are coming out of the transformer, for some reason I am not able to show that here. Now, the requirement is if any error occurs while loading records, there sho...

Same pearl with a different shine. You'll probably need to use user-defined sql. :wink: First, clarify what 'col2 < firstday_of_month' means. Col2 is a date? I doubt you can use 'MM' here without also including the year as well. I'm guessing you need a sql statement more akin to this: delete from t...

I'd suggest stepping back and not using user-defined sql, it's not needed here. Create one column in the stage for col1 in the table and mark it as a Key field. Use an Update action of 'Delete existing rows only'. Set the value of the key field passed down from the Transfomer to 1. Done. OK that wo...

That was just to get your job compiling and running with a chance of success. The actual setup of the DRS stage for this is a completely different issue. Post your delete SQL. The sql is realy simple. delete from tablename where col1=1 In the DRS stage, I have specified the tabel name in the "...

You need to define a stage variable so the job will compile, no need to actually use it. Then a constraint to ensure that only one record goes to the database stage, otherwise the job will run forver. I use: @OUTROWNUM=1 I tried this. 1. I tried with defining the table name in DRS. Also I defined o...