DSXchange

So basically the number of splits should be dynamic. What column in the source record will tell how many records need to be created? Can you please give two proper examples with data. How the incoming record will look like and on what basis it will be split and on what basis to decide what column go...

By parametrizing the names.

Well then give us the complete requirement. Looking at your sample data, it tells me two things. 1)Split a record into two, and 2)First three columns in one record and first and last column in second record. Now your telling me there are going to be multiple records Please provide a full description...

How many records are we looking at? Cant a single instance take care of the load? Use different names for different instances for the control file. Is SP.txt the data file?

Refresh you database and check the length again. Reimport the table definition. Try again.

YOu can do it a couple of different ways. One way would be to read the file twice using two different sequential file stages, stick a transformer after them. Choose Col1, Col2 and Col3 for the first stream and Col1 and Col4 from the second stream. You can then write it a sequential file (1 from each...

Also takes the load off your database server.

chulett wrote:Thank goodness.

Make sure you run in sequential mode or the results might be different.

Did you try Ray's test?

As mentioned, no px routines are shipped with the product, but functions are available. If you want to use server routines in a px job, use a basic transformer.

Pay attention to ds_develper's suggestion. A very valid point.

ray.wurlod wrote:
Question: what if NLS is enabled?

Good question, will have to try that out some time.

Thanks to Craig who came back and shared the resolution with us in the embedded post.

Wait for the TX folks to swing by with some pointers. In the mean time, I suggest going through the manuals that are shipped with the product.