Count Jobs in a project (Parallel, Server, Sequencer)

[kmsekhar]

Hi All,

I have jobs names like:

Server Jobs - ser_Sample1
Parallel Jobs - parl_Sample1
Sequencer - seq_Sample1

generally I am using the below commands to get the count

Code: Select all

To Count Parallel Jobs in a Project
SELECT COUNT(*) FMT '60L' FROM DS_JOBS WHERE NAME LIKE 'parl_%';

To Count Server Jobs in a Project
SELECT COUNT(*) FMT '60L' FROM DS_JOBS WHERE NAME LIKE 'ser_%';

To Count Job Sequencers in a Project
SELECT COUNT(*) FMT '60L' FROM DS_JOBS WHERE NAME LIKE 'seq_%';

Is there any alternative to get the count individually (Parallel, Server, Sequencers)?

[chulett]

DS_JOBS stores the job type in a field called JOBTYPEIND from what I recall so you shouldn't need to rely on a naming convention. Search the forums for the various values and caveats on its use.

[ray.wurlod]

Or don't even worry about the values, just report them.

Code: Select all

SELECT JOBTYPEIND, COUNT(*) AS JOB_COUNT FROM DS_JOBS WHERE @ID NOT LIKE '\\%' AND @ID NOT LIKE '//%' GROUP BY JOBTYPEIND;

[qt_ky]

An alternative would be to use the command line dsjob -ljobs syntax piped through the grep command based on your naming convention, like grep ^seq, and then pipe that through the wc -l command to get the count. Try it and see if it works or if you have any questions.

[chulett]

Or don't even worry about the values, just report them.

Unless you're not sure which is which.

[pandeesh]

The below are the valid values for JOBTYPEIND:

Code: Select all

0 = server job
1 = mainframe job
2 = job sequence 
3 = parallel job

[kmsekhar]

Thanks all

[PhilHibbs]

Unless you're not sure which is which.

It should be obvious! If I saw numbers 537, 86, and 2 I'd know that for my current project that those mean Parallel, Sequence, and Server respectively. Sure, for some projects there might be some ambiguity but I'd guess that that is very unusual, one job type will usually dominate the stats.

p.s. another way to get these stats is to use my DataStage Analysis Spreadsheet to analyse a DSX export. It's a bit slow processing exports with lots of parallel jobs, but I've got a fix in the works for that.

[chulett]

One man's obvious is another man's wtf.

Sure, it "should be" (hence the smiley in my post) but honestly, why would I want to report on how many "zeroes, ones and twos" each project had? Was just suggesting that the OP take a moment to lookup what each value means so there's no ambiguity no matter who sees the counts.