DSXchange

Here is SQL to find a hash file. I assume you can modify to find a job. select DS_JOBS.NAME AS JOB_NAME FMT '35L', DS_JOBS.CATEGORY, DS_JOBOBJECTS.NAME AS LINK_NAME FMT '35L' from DS_JOBOBJECTS, DS_JOBS where DS_JOBOBJECTS.OLETYPE in ('CHashedOutput','CHashedInput') and DS_JOBOBJECTS.OBJIDNO...

Nice word doofosity. Latest buz word?

EtlStats works on Windows as well. If you have any problems installing then just let me know.

I posted a series of jobs on ADN and on my tips page which parse the XML version of these reports into either hash files or Oracle tables. You can download EtlStats.zip and modify these jobs to do whatever you want with this data. There are several posts on how to use these jobs. Do a search to see ...

You will need to create the tables. The scripts are there. You need to run BuildJobStatusHash. This builds a hahs file used by DSJobReportDb. The controlling job is DSJobReportDbDriver. If you put All in the Category parameter then it will load row counts for all jobs. To get row counts at the end o...

We run this at the end of each sequence. It posts start and end times, duration in seconds, rows and rows per second. It also posts parameter values. All of these can be used to measure performance over time. It also allows you to predict when all your sequences will finish. Soon I will have bytes p...

That may be true as far as a hack. I think it is safe to use.

This only works in a sequence job. You have to run it in a routine activity. I doubt if it would work in an ordinary job.

I still do not like this method. I think it is cleaner to leave the result in a hash file and use the routine posted to change the parameter value to the one written to the hash file. I can do one or a hundred parameter values this way. All I need is a temporary holding place like a hash file or a s...

You could download EtlStats.zip from ADN or my tips page. It posts all the row counts to tables in the warehouse but you could easily write these to a text file. You get all the jobs and create table statements. It posts link names and row counts. Very simple to setup. It will also do this for all j...

I like the SQL solutions. I do not like the API solutions for logs and row counts. Row counts are easiest with dsjob -report. IMHO. The API works best for starting and stopping jobs. I just wrote a routine which will change these SQL statements into html. I will post it soon. I want to tie this to a...

A sequence job and executes commands, routines or just runs jobs. It has no input or output.

You can run a different sequence from within a sequence if that is your question. Most posts call this a child sequence. A sequence is just another job as far as DataStage is concerned.

Thanks for the good words but I am not sure how this is working now. If this is in a batch job or a sequence then it may be a little different. Usually I have a parameter which reflects the invocation id either exactly or logically. When I get row stats then the main invocation id is the job sequenc...

I would put everything you need to check into a hash file. Make every column a key field. The only field that is not a part of the key is the surrogate key. We call this a Xref hash file. Hash files are normally rebuilt every run. We select the max(surrogate_key) using a UV stage from this hash file...

Look at http://www.dsxchange.com/viewtopic.php?t=89892 as well.