DSXchange

You need to do more than change the datatype in the job, you need to do something like CAST on the field in your source query as well.

As noted, try casting the column to a supported datatype. Or upgrade DataStage.

@chulett : Have tried it through full outer join in Join stage and it worked. previously i thought it won't work and posted for suggestions to do it in some other way.... All good information to have included in your original post. Without it it just looks like an interview question. @stuartjvnorto...

BrunoD wrote:This parallel task used to work fine.

However, you haven't told us what is happening, what exactly is no longer working fine? What error are you seeing?

What have you tried?

Are you certain that your stage variable always contains 6 numeric digits?

it seems like RT_LOG2451 file doesn't have the write permission. am i correct? Yes, for whatever user you are using that is correct. Use the proper user, the one that does have the permissions it needs. As to the "this one too" error, you don't do any log clearing while logged into the UV...

ray.wurlod wrote:Disk is cheap these days.

Pfffffttt... not on a government owned server.

Good to know.

Override the default assigned auto-purge settings for this job to keep whatever you need it to keep to figure out the issue.

Set it as the value of the Project level environment variable.

SURA wrote:The loop is not going to wait for the existing job to finish!

So this isn't true. If it is using the syntax you posted, it is in fact waiting for the job to complete before moving on to the next line of code.

Of course, but I assumed the question was can that all be pitched and replaced with strictly 'parallel' components... at least that's where my response was based on.

True but I think you actually have to search for something, as in you cannot leave the search field empty and get everything.

There is a crap-ton of stuff under the covers that the 'KBA Utilities' encompasses... and I really don't think you could do anything equivalent as a straight parallel implementation. Even with a couple of custom stage types and/or routines.