DSXchange

Why complicate it? Just use Count() or perhaps DCount() although that would change what I posted... slightly.

Count the number of "_" delimiters. Take the "count+1" field for the date and the "1 to count fields" for the file id. For your example, count = 5 so field 6 and fields 1-5.

Amuktamalyada wrote:I am essentially using one lookup stage instead of many lookup stages

That makes absolutely no difference.

knnath wrote:If generating a flat file and importing it to mdb format, do we need some drivers to be installed?

No. And technically, I meant "import it into Access".

Again... dot slash.

And just to state the obvious, since you don't have the current directory in your path, make sure you prefix dsjob with "dot slash" if you park yourself in the engine's bin directory:
Or you could fully path the command. All this, UNIX Command Line 101.

So... resolved?

Seems to me you mean "real life" rather than "real time"...

MDB is the old format for Access database files. I would assume you would either need to obtain ODBC drivers for Access, that or just produce a flat file and import it.

Only parallel jobs? Modify, yes, Transformer... no.

Look in uvconfig to see if you have any NLS settings (like NLSMODE) mentioned. No entries = NLS not installed. Also note that NLS can be installed but not enabled.

ps. You shouldn't need NLS to convert EBCDIC to ASCII... depending on the target characterset, I suppose.

I don't see why not - post your derivations. If they use Convert() that is available for both Server and PX jobs.

It might be helpful to know what's in the BASIC transformer...

( found that information here)

1)Define a stage variable If Isvalid ("date", Datecol,"%yyyy%mm%dd") then 1 else 0-sv 2)Use constraint sv=0 A couple of learning points: There really isn't any difference between the two posted approaches (other than one checks for true and the other false), so unless we know th...

Surely, someone does. Can you not request that they do this for you?