DSXchange

I you are starting this job yourself in the script, why not just add "-wait" to the dsjob command and avoid the loop?

I'm not at a site where I can check the query syntax, so am hesitant to post SQL that might break things. Can you re-export with the correct READONLY settings and then import it into the project? p.s. Please don't use all CAPITALS, that is considered to be shouting, and there is no need to do that y...

I cannot think of a scenario where a record could get rejected where the job should not get aborted (i.e. disk full, I/O system error, etc.)

The is no column limitation close to 80 for DataSets. If you change your job to write to a sequential file (pathed to /dev/null) does it run correctly? Where are the data files written to (specified in the APT_CONFIG_FILE) and is that/are those disk(s) filling up?

That is correct, the DataSet has no reject links. What kind of a condition can you imagine that would cause a reject (as opposed to a job abort)?

There is not stage that can perform mathematic operations on non-mathematic objects. You need to define what the result of "B+Z" is, as BugFree has noted. Once you have a rule for that you can either convert your values to numerics and use an Aggregator stage or you can write your own code to do this.

The sequential files will FTP without a problem, but the hashed files will not. You can copy them over and then convert them using the FNUXI or FORMAT.CONV commands.

No procedure necessary. Import into Version 8 and recompile; do a test run with known data to compare to your original V7 results and if they are not the same you have a problem.

It makes sense to use UNIX path notation on windows as well so that you can cross platforms without having to convert.

It would seem that there is ECase 118247 open on this issue and is states

(contray[sic] to the documentaion[sic]) that IsValid only accepts date format %yyyy-%mm-%dd

. I am pursuing this, as I don't think that is acceptable and will update this thread when it gets resolved one way or the other.

Try adding "-mode NORMAL"

Those are user-written functions - look at what they do in server

ashishpascal - there are too many ways to slow down and speed up jobs to be able to answer your question. First step - know thyself. Is your read or write the bottleneck?

consulting wrote:...dslinkname=11 during compilation itself it fails...

With what error?

No. If the value you have is a string you need to convert it to date.

That looks like a user-defined function, so without knowing what it does there cannot be any suggestions as to how to do this in PX. Look at the function or perhaps post it to this thread.