DSXchange

Hi,

Wanted to know the best approach to a complex problem.
I am expected to flow data from 2 flat files into a huge XSD( XML format which is existing ) in step 1 and flow data from XML stage to RDBMS ( Informix)

my flow is like this:

File 1 and 2 ---> Transformer ---> XML Output (Step 1)

Where I use hash files or seq files at the left side of the transfomer and map the relevant columns to the xsd format which I have imported(this is on the left)

Now for step 2

XML Output -->XML Transformer--> XML Input --->Transformer-->RDBMS tables

I have found that this is not working satisfactorily.

Manuals say that XML input /ouput are to be used ...is there a better way. Can someone comment on their prior experience?
My main stick ups are

1. The file(s) have a number of nulls and missing data
2. The XSD is huge with atleast 70 tables and further column names
3. Is the process route which I am taking correct?
4. Can I use the same for PX?
Has anyone worked with PX and XML?

As I am scheduled for a rush delivery, will be glad if some one can help out!!!

Regards
Srini

Gurus:

Could you help out on the server side?

Regards
Srinivas.B

Srini,
I can't help from asking why aren't you populating the Informix directly ? (you must have a very good reason to do it that way - I mean construct the XML and then de-constructing it for the RDBMS population)
Let's say you need to deliver the XML & the RDBMS - then do it on parallel

job 1-File 1 and 2 ---> Transformer ---> XML Output
job 2-File 1 and 2 ---> Transformer ---> RDBMS tables

Now for some tip on how to encode XML with DataStage see

viewtopic.php?t=85500&start=0&postdays= ... highlight=