DSXchange

Hi,

You describe a known bug. If you use a function in a user defined sql, decimal are seen as (38,10), and string fields are also with a length fixed by Datastage (I can't remember the length).
Just type your decimal as (38,10) and change when possible.

Matthieu

Hi,

We had the same issue with the the Oracle Loader with Datastage 7.5.2 (it was working well in 6 version) and looked for a solution. The IBM answer was : "use job handler" (!!!).

Bye
Matthieu

Hi durgaps , As you were thinking about a permission problem, I asked the way you were running DS, because depending on this, the profile loaded can change. When you run DS in command line with your profile (witch is not dsadm), jobs are running with the dsadm environment variables. So you may need ...

Hi,

How do you run your job : debug, director or command line ?

Matthieu

Hi alone,

The first character is reserved for the sign... For example -0490.0.
This is not a problem with ODBC connection nether OCI plug in. If you store this decimal in a sequential file, use a CHAR/VARCHAR type.

Matthieu.

Right !

Matthieu

Hi, If your are working with a "fixed width column" flat file you can also use this unix command line : cut -cx-y file | sort -u | wc -l > tempfile x : starting character y : ending character file : your file's name (better with the path) tempfile : file to store the number of groups (bett...

Hi,

I think I would sort and "Remove Duplicates" (partitioned) to keep the distinct group values, and then count (sequential) the number of groups with @inrownum or an aggregator...

Matthieu.

Ok,

So, just create a flat file with 40 lines containing "1"...

Matthieu

Hi, What is the MF edition ? What a strange idea !!! You can use a row generator to create 40 lines with one column (called "JOIN" for example) filled with a fixed value (for example "1"). Then you create a column (also called "JOIN") in your "stream" flow wit...

Hi,

Why don't you use a UNIX command in a before job subroutine ?
tail +5 yourfile > newfile

Bye
Matthieu

Hi, The two big good points with my method are : - Your file can be "unstructured". It means you can have a variable number of fields. For example : 1,2,3,4 4,5 8,10,20,21,12,12,14,21 - You don't have to modify your developpment if some fields have to be added. That's a good point isn't it...

Hi, May be something better for you would be to replace "," by a end of line (I think it is char(013)). You need first to flag each line individualy with Surrogate_Key_Generator stage, then you obtain a file like this : Surrogate CodeFile 1 1 1 2 1 3 2 2 2 3 2 4 Instead of CodeFile 1,2,3 2...

Thanks Ray, but the Modify stage is not available in version 6 of DataStage PX...

Matthieu

Thanks all, ucf007 : the number of left zeroes is variable. It would be too simple with a fixed number ! pnchowdary : FMT is a server fonction and the basic transformer is not available in version 6. ashwin2912 : I thought of this solution, but I don't like it much. If I don't find another one I'll ...