DSXchange

You could do this lookup in a routine. You can easily count the multivalues and do the lookup in one read.

I does.

_________________
Craig Hulett

You does. You promote all jobs as a release from test to prod?

No you can include these types of files in Version Control. These files have nothing to do with compiling.

I don't anyone promotes all jobs as a release. I don't trust VC that much. Everything is just done as a version to a single job. VC is limited in how much it can do.

Also VC needs its own project. Do not use Dev as your VC project.

I cannot help you if you do not give me more information. This link does work. When you build your lookup hash file then you have to write immediately because you are going to lookup on the same hash file. I would also sort the incoming data. Lets say you are looking up a primary care physician for ...

You either have to know the data or write something like ProfileStage for hash files. If you count how any fields have multivalues and build a list then you can figure out the groupings. Now a field maybe part of a group if it is null. A group should all have the same number of multivalues. So lets ...

I have done this but it does not immediately show up in the Director. I am not sure why. But clearing right records in RT_STATUSnnn is fairly easy. The RT_LOGnnn is little more compilcated. I am not sure but I think they write something into DS_AUDIT as well to force Director to reread the RT_LOGnnn.

You need to explain your need in greater detail because the solution pointed to will create a multivalue field by updating the a hash file with the new value added as a multivalue to the same field y doing a lookup on the hash file. You need some logic to pick the correct multivalue but it works. So...

The instance ids are stored in 2 places in RT_LOGnnn and RT_STATUSnnn. You could hack these and delete these directly as Roy stated but I would not recommend that rither. You can setup autopurge and delete anything over 2 days old. I think autopurge with reusing the same instance ids is the best ide...

The only other thing I can think of is are your copying from a server job to a server job? Are the jobs the same type? I doubt if you can copy from a server job to a sequence or to a PX job.

Should work. Are these in one Designer and not two different connections?

I think Ray is correct. Do full loads or normal loads. The row counts should match. Sample data is much harder to create and does not prove much.

That is not an aggregator solution. It does aggregate. Cool huh?

If you want to try to put something to together then KgdDsJobView routine can pull out the metadata from a job defintion. The routine is here m You need to change it quit a bit to use it as a routine. The SELECT needs to have "TO 9" on the end. It may not let you open DS_JOBS or DS_JOBOBECTS directl...

MCM also switches the registry entries.