Running PX jobs for a limited number of rows

[satish_valavala]

Hi,
I want to run a PX job for a specified number rows.
But, in Job Run Options --->Limits Tab --->Rows -----> both the radio buttons No Limit and Stop Stages after Rows options were disabled.
Could anyone tell a way to do it.

[richdhan]

Hi Sathish,

All these options are enabled for a server job but not for parallel job. What do you mean by limited number of rows? Do you want to run for the first 50 rows or you want to limit the number of rows based on a condition. In either case you have to use the Transformer constraint.

HTH
Rich

[satish_valavala]

Thanks Rich.
I want to run the job for 50 or 100 or 1000 .....rows
Is there any otherway to do it( without modifying the job)

[kcshankar]

Hi,
We can use Head Stage.

[Madhav_M]

In case of server jobs you can invoke the job using

$DSHOME/bin/dsjob -run -mode NORMAL -rows n <project_name> <job_name>

[roy]

Hi,
I don't exactly understand your reason for wanting this but it depends on what is your source, here are 2 examples:
1. A sequential file has a "read first rows" option for reading n number of rows (you can use a parameter as value)
2. An Oracle stage can use a where part constraining n rows (also use the $DB_ENABLE_RESERVED_CHAR_CONVERT to allow parameters in the where part to be translated by DS); I think you can't get a specific n rows returned this way since it might not be the same n rows each time.

IHTH,

[aesguerra]

Hi,

If I remember correctly, there is a tab in the job properties window where you can specifiy the number of rows that will be processed. The rows that are processed are then printed on the log file similar to placing a peek stage.

[sun rays]

This option is enabled in server jobs, but not in PX.
Also you can do a work around as Roy said, but this would only put a limit on the input rows and not on output rows. For Ex your job may be limited to read only 1000 rows and process them, but your output may still have just 10 rows, or 2000 rows.