DSXchange

mv FILE1.txt FILE1.txt2 gives us the impression that you are just renaming the file extension, unless FILE1 is different for both the files. Regardless, you are renaming the file. Provide fully qualified paths. Something like #!/usr/bin/ksh export FILE1 = "fully/qualified/path/of/FILE1.txt"...

Do this: Field(in.Col, "$",1):Field(in.Col,"$",2)[1,1]:Field(CONVERT('ABCDEFGHIJKLMNOPQRSTUVWXYZ','',Upcase(in.Col)),"$",2) Get everything before the dollar sign, concatenate it to the first character of everything after the $ sig...

Have you gone through the DataStage Plug-In and Installation Guide. I also recommend the Parallel Job Developer's Guide.

"They" might come out with another dot releases for vista compatability. Have not heard anything so far as of yet.

Also make sure directory &PH& is kept in check as that can reciprocate with slower response times. And basically clear the status file and keep the log file in check as well.

Get 7.5x2 if you have to use parallel jobs. As simple as that.

kumar_s wrote:DSguru2B, atleast for this sake you can display you name DSguru2B .

Name, o yea sure, its Brian

Use

venkycool wrote:Ohhhhhh...is it? Thanks Brian and it was very informatiove...will keep this in mind...

O yea. Thank Brian

Hmm, ok, I will have to test out a small routine written in C++. Thanks for the info. And yes I am Brian . I guess I have to take out Brian's name.

Vincent has an FAQ entry on the same.

This happened becuase of CR. DOS files have CRLF as line terminators and UNIX files have only LF. Since it was a DOS file and specified as UNIX, it took LF from the CRLF and left out the CR which came in as thickened pipes.

Usually two timestamps are stored in a control table. These timestamps are passed as a job parameter to the sql select (select * from table where date between #date1# AND #date2#). Once your cycle finishes OR before your next run, the control table is updated/inserted with new date records and selec...

No problemo. That comment of mine was a hasty move. I got stuck with some junk data in my output and I jumped to that conclusion. No wonder Ray calls me a grasshopper.

bcarlson wrote:I think a buildop would be much easier. Then there are also no restrictions on sequential vs. parallel. That always throws a red flag for me.

No restrictions for sequential vs. parallel in a routine as well. I think it depends upon how a person wants to do it.