DSXchange

Check with your UNIX administrator. Some UNIX operating systems must have large file support explicitly switched on. This may not have occurred, which would explain the 2GB limit.

You don't seem to understand what the Folder stage reads. It has two "columns" - the first gets the file name (which must be non null, for obvious reasons), the second gets the entire file contents as a string. This is not the correct stage type for what you want to do. As Craig suggests, an operati...

Arnd actually meant SEQ(<1char substring>) to report the ASCII code point of a character.

Do you have NLS enabled? This will affect the final answer. You should be able to handle accented characters without, but how you do it depends on whether it's enabled or not.

Welcome aboard! :D What were your expected and actual results? (You can use stage tracing to determine these, capturing them into a file, or the Debugger to see the values either side of the Transformer stage.) You can post the Routine code here if you like but, if you do, make sure to include the F...

No, it means simply what it says - you have had an activity abort and have not provided any handling for it.

"Handling" might be a "Failed" trigger, some other trigger whose condition includes aborts, or an Exception Handler.

Use an Execute Command activity with rm as the command. The command arguments field can contain options (for example -fr) and a job parameter reference to the parameter containing the name of the file to be deleted.

Ken's response shows you the standard technique; you need to determine whether the key already exists in the target table, and on that basis make the decision whether to insert (a new row) or update (an existing row). Life becomes marginally more complex implementing Type 2 slowly changing dimension...

This is one of a series of posts where sudheerkumar is experiencing difficulty installing (a Windows version?) of DataStage on an unsupported Linux platform (SuSE 9). Until we get a statement that it is not a Windows version being installed, there's not a lot of value trying to pursue why shared mem...

Surrogate keys are ONLY about uniqueness. There is no requirement that they form an unbroken sequence. If there are gaps, this does not detract from the fact that each is unique. Change your mindset.

Click on the GetSQLInfo button (stage properties) to ensure that your query is constructed with the correct quote and delimiter characters. When you're building up to it, as Arnd suggested, test with a column name that has no spaces in it before introducing a column name that does have spaces. It do...

Your reported results are probably just a statistical artifact of rounding to whole seconds, and doing so before performing the division for rows/sec. For example, a stage/job that completes in under 0.5 seconds will report 0 rows/sec (which technically means no rows processed, but the row count rep...

I think you'll find that the Oracle DBA has privilege to do ANYTHING - that's the definition of a DBA!

The easiest way to create one is right-click on the warning message itself. One of the menu options allows you to create a message handler for that job for that message.

I'm not on 7.5 right now, so can't answer your second question about where the message handlers are stored.

Startup time includes all the time needed to get the conductor and section leader processes started and may - I'm not 100% sure on this - include time for the job to reach a status of Running on all nodes. Production run time is that period between when the first operator begins and the final operat...

Try just putting sh (or /bin/sh) in as the hard coded part.

On the parameters line, add a job parameter reference to the pathname of your script.