DSXchange

Do all the tables have the same DDL? If so, then two methods come to mind immediately: Read the file within a sequence, looping once for each table, then (a) call a PX job passing the table name and use the appropriate write table stage for your database and use the "create table" option. ...

I wouldn't hold my breath if I were you. This has been around since local containers were first implemented.

Does your "TempSource" DDL has a "bigint" as the column datatype? If so, then it might be that the ODBC driver converts this to a DECIMAL(19,0); in which case you can either declare the variable as a decimal(19,0) in the datastage input, or turn off the warning and let DataStage ...

Add the "microseconds" to your datastage timestamp column. If that doesn't work, you need to shut off "fail on size mismatch" in the ODBC stage. Do you want or need the 3 microseconds? If not, then do a CAST in sql to change the type to one with no microseconds. The sytnax depend...

No, you've discovered that this functionality doesn't work. I usually end up making a copy of the job, unconstructing the container and running that copy to see the designer statistics.

The "folders" shown in the director are just virtual attributes to jobs, thus one cannot simple rename a folder but one needs to do as Craig has already suggested.

The error concerns the precision of microseconds. The timestamp coming into datastage is "2012-09-10 12:10:12.567" while the datastage type isn't declared with any microseconds; and since schema reconciliation in your stage is turned on (you can turn it off) the error message is displayed....

As long as the source and target mappings are known then using NLS DataStage will not be an issue and it will correctly perform mappings/conversions.

Are you sure that you are picking up all job types and doing a "force" compile?

What happens if, in the designer, you click on the job name, do a right-click and mutiple-job-compile just that one job?

yes.

The only way to call server routines from a parallel job is to use a BASIC transform stage and call them from that stage. The exceptions are the before- and after- job routines which are BASIC ones in all job types. Parallel job routines are calls to objects which use c++ type variables and are typi...

I can't reproduce the problem on my Windows 8.7 machine, "StringToDecimal('-77')" into a Decimal(10) field keeps the negative sign.

Yes, it needs to be compiled and catalogued.
Yes, you can compile it - but first you need to find the source.
You will need to do a "CREATE.FILE BP 19" to create a BP subdirectory in your project, then copy the source into it and the compile as suggested above.

The before-job gets called before any stages in the job, so container usage is unlikely to affect this problem.

I meant to look at the source code of your "DeleteHashFiles" routine - what happens between the start and the first DSLogInfo() call that might terminate it?

Do you realize that you can put a reject row into a transform stage and then add/delete/change columns there? Do we have all the basic functions, link functions available in basic transformer in parallel transformer? If not, in the reject link of any stages in parallel job can we add few other colum...