DSXchange

I missed the CHAR definition in Oracle in your original post. This datatype has no terminator, as it is fixed length. Normally DataStage will pad this column with spaces, but in order to do your queryto function you will need to use another pad character, the null (0x00) character. You can use the t...

Using LOAD invokes the Oracle bulk loader and uses a different load technology. The bulk loader will use $APT_ORACLE_PRESERVE_BLANKS setting, plus you can also use the data type VARCHAR2 which will also trim trailing spaces automatically. You might also change your job to do an explicit TRIM() on th...

If you have a client installed then you do have the documentation; if you don't have a client or a server then any explanation here is going to be somewhat useless. There is no reason to make an inclusive list of active and passive stages since they will change over time and with addition of package...

a dsapi_slave process is forked from the dsapi_server on the DataStage server machine in order to do actual processing of a datastage client. There is little to explain.

Yes, the dsapi_slave is a datastage process.

If you right-mouse click on a column in the input sequential stage and go to "edit row" you will get a window that lets you set many different attributes on a per-column basis. Click on the string attributes at the lower half of the window to see your options, then you can use the online help and/or...

There is a bug in 7.x which causes a process that attempts a DSAttachJob() to itself will "hang" on a semaphore for 30 minutes before returning an error. I've changed all my code to check for JobName = DSGetJobInfo(DSJ.ME,DSJ.JOBNAME) before attempting a DSAttachJob(). This might have been your prob...

It looks to me as if the problem doesn't lie with your particular job, but with the installation. If you write another job with just a row-generator and a peek stage does that work? The line in the error message stating that the process cannot attach to "/app/dstage/Ascential/DataStage/DSEngine/Proj...

Within DS I think you will have to use the TRIM() function in a transform stage to do this. There are string handling capabilities in the sequential file stage, but they won't work for all types of strings (for instance no implicit trimming will be done when reading quoted strings withi embedded tra...

Vincent,

I'll give that a try tomorrow and also await Ascential support's response. I have trouble believing that they would allow a repartition to DB/2 without letting the person running choose a user-id/password; particularly in cases where the database is on a different node.

I just realized that you need to specify "-64BIT" on the command line; using jsut "64BIT" will not result in a 64bit file (but it won't give you an error message, either). Sorry about that.

The "connection" error in the message is between the client and server. If you can read/write/compile the same job in your Designer then the problem is most likely not due to problems on the server side. Can you start other jobs from your Director? If not, de-install your client completely (making s...

Just define a third column that contains the rest of the data in your fixed-length data and don't use the column. PX needs to know what it should do with the extraneous data and this is one way of handling that.

The way to do this is to use the UNIX command that actually creates the file: mkdbfile {filepath & name} 2 1 4 64BIT this creates a type 2 file with a modulo of 1, separation of 4 and 64BIT pointers. You can use the hashed file stage and file creation command in a job, run it and see the actual ...

Vincent, I just went through all the screens in the Transform stage that I could find to no avail - I cannot find any window where the userid or password for DB/2 repartitioning can be entered. This job has no DB/2 stages. Nothing visible to me in the Job properties or elsewhere. Could you perhaps p...