DSXchange

What is your delimiter character? Does this character appear anywhere in the data? How, precisely, is field6 defined? Is what you've posted on one line or three lines in the source file?

No.

The only, repeat only, version of DataStage lower than 8 that supports parallel execution on a Windows platform is 7.5x2

It can be done without the TRANS() function using a "proper" join. SELECT DS_JOBS.NAME, EVAL DS_JOBOBJECTS.'@RECORD<31>' FMT '32L' COL.HDG 'Dependent Job(s)' FROM DS_JOBS, DS_JOBOBJECTS WHERE DS_JOBS.NAME = '<<SequenceName>>' AND DS_JOBOBJECTS.OBJIDNO = DS_JOBS.JOBNO AND DS_JOBOBJECTS.OBJTYP...

System tables. You don't create system tables.

You might most easily be able to accomplish this in a shell script invoked from an External Filter stage. There does not have to be a one-to-one mapping between input and output row counts from this stage type.

I'd be inclined to create a routine to be invoked from a Routine activity in which the list of addresses is read and composed into the Addressee field in the DSSendMail() function call.

Is that Varchar2(50 CHAR) is equal to Varchar(200) in DS because of NLS_CHARCTERSET=ALT32UTF8 settings in Oracle?? No. That might give 100, but not 200. If I read it as Varchar(200) and write to sequential file as Varchar(50)- then is there any possibility for truncation of data?? Yes. Think about ...

Try using the Ascential-supplied ODBC text driver.

And/or make sure that the ODBC DSN is configured with the correct format option (e.g. CSV, first line is header, etc.)

1024 / rowsize

(Array size * row size) should be close to, but not over, network packet size (or a multiple thereof) for optimum transmission. (Rows per commit) - or transaction size - is mediated by a number of factors, but should always be a multiple of array size. Array size of 1 with rows per commit of 1 guara...

Prevention is better than cure.

Make sure that the T30FILE tunable is large enough for the maximum number of open hashed files system wide at any time.

Monitor your hashed files and resize them to use 64-bit addressing if they go over, say, 1.5GB.

Re-compile the job sequence(s) giving the problem.

Do you have any bitmap indexes? These are notoriously slow to update.

Is this some new scheduler out of the deep South, or just a typo?

"Scientific notation" in DataStage requires an upper-case E. Thus 2.345e-006 would be regarded as a string, but 2.345E-006 would be interpreted as a number. You can use the Num() function to verify. The maximum supported exponent in DataStage is +38, the minimum supported exponent is -38, and the ov...