DSXchange

That's because a single instance job can not be started if its status is already running.
You will need to give each instance a separate invocation ID - perhaps the name of its controlling job sequence (which can be obtained from the DSJobController macro).

Systematic debugging is a vital skill if you're thinking of a career in IT.

In a word, AFTERWARDS. Trying to capture these statistics in real time does nothing but slow down your throughput. Further, it's probably something you don't even want DataStage to do. Number of records in source is a property of the source. Number of records in target is a property of the target. U...

Show me. The script looks OK, so it has to be tested thoroughly. Show me the results of a script with echo !$variable! statements in it. The "!" characters let you see where the variable is empty. I presume a manually entered dsjob -logsum request with no variables works OK? Then start adding variab...

1) No. There might be something in the ODBC driver, depending on the vendor from which you obtained the driver. 2) There's no automatic one, nor would it be easy to detect a failed channel part way through. You could detect the specific error message in the log and built your restart logic around th...

Debug it yourself.

Prove that the script variables jobname and jobstatus are being evaluated correctly. Prove that the environment variable PROJECT has a correct value when this script is being executed.

(Use echo to stdout as one easy way.)

Ray do you happen to know of the reviews on advanced EE training from IBM? Student feedback forms are returned to IBM. After that they are analyzed to look for ways to improve the classes. The instructor quickly looks at them (maybe) at the end of class, and never sees them again. If there are issu...

Since SDKSequences is a hashed file, navigation is by field number, not by column name. Create a table definition containing two columns (called, maybe, SequenceName and NextKeyValue), both VarChar with precision 254. Mark the first of these as Key and not null. Use this table definition in Hashed F...

The "#" delimiters are required in passive stages. The are not required anywhere else.

They are file names. They can be as long as you wish (up to 255 characters). These files contain operating system group names. As far as I am aware there is no eight-character length restriction on these either. Execute the command cat /etc/group to determine the group names in use on your server.

This is not something that you do in DataStage. You do it when creating the ODBC data source name using odbcad32.exe.

If you're reading the flat file with a Sequential File stage why not just specify sort -u (with appropriate redirection) in a Filter? Then process the file that is the result of that command, and which will have no duplicate rows. I am assuming here that you have some kind of UNIX-on-Windows, such a...

I have tried to execute one procedure in stage ODBC: User-defined SQL=call nameprocedure(); The procedure is executed several times (loop). How I can avoid the loop? Thanks The stored procedure is called once for each row processed by the job. The only way you can "avoid the loop" is by processing ...

All row counts are reported in "active stage finishing" messages in the job log. You can view these in Director, or access them programmatically using functions such as DSGetLogSummary(), DSGetNewestLogId() and DSGetLogEntry().

Client licensing is now being enforced, where it wasn't prior to version 7. Licensing is checked in three different locations, only two of which can be "hacked", so forget about trying to defeat it. There can be a problem of licences being held by defunct (abnormally terminated or killed) connection...