DSXchange

yes, you can write a small routine that reads in a record, duplicates it 20 times and cats it together with a LF in between to split it up. Something like ToDup=trimB(Arg1) For n = 1 To 20 Ans=:ToDup:char(010):char(013) Next n You can enhance it to fit your requirement. But t...

Do an ereplace of that character with char(010):char(013).

Look into the link collector. Have twenty links coming out of the transformer and collect them with the link collector. This way you will have 20 copies of a single record.

yea, you will have to code it. Your routine should be smart enough to handle leap years, and know when to add 30 days and when to add 31 days. Its not easy, i agree, but its dooable. OR if you want the easy way out, load your flat file in a temp table and then use oracle's function of add_months in ...

I was under the impression that multiple files need to be unzipped and hence i advised to go for a routine where a substantial amount of error handling can be coded. But if the unzipping is for a single file then yes, no need for a routine.
Regards,

Instead of making a call to oracle in a routine, just custom build the routine to add 30 days or 31 days to the iconv of the date. Search this forum for many examples.
Once that is done, you can do further manipulation in the transformer itself.

Did you try putting the end value as a negative number?

Search is your best friend around here. Check out the last reply in thispost.

You can also use the routine activity stage in a sequence job to call the custom routine that unzips the files. This way you can control the execution incase the unzipping is unsuccessful.

For that you need to get your data modeler/data architect involved. Lay out a rough sketch of how the data is combinable. Handle cons, create extra tables for different/new dimensions, create fact/factless fact tables, or add new attributes to the existing tables.

Is the module already a working piece???
Well, make sure you have enough disk space for accomodating staging files/reject files. Also, make sure you dont create any file in the datastage directory.
My 2 cents.

That doesnt look like db2 timestamp to me. Isnt it suppose to be
YYYY-MM-DD HH:MM:SS.nnnnnn. That makes up 26 bytes.

or you can use the following code

OR

I did something like this a few months back. Check out thispost for more details.

Google for sed operation. I think you will need to use something like

You need to double check my syntax.