DSXchange

I came to know from the support team that dsoper has full control over the infiles folder. Any ideas? Don't believe anything you're told. Demand proof (for example a long form dir listing). The symptoms suggest an inability to read the file because it can't be found. This might be permissions in an...

Symptoms? What were the most recent events logged? Have you got any progress reporting so that you can pinpoint the point of "failure" more accurately? My advice would be to put some in - maybe just before-stage and after-stage subroutines that report "got this far - in stage xyz".

You need to construct pairs of stage variables, or calls for each of the key columns. The problem with RowProcGetPreviousValue is that it can only deal with one key column. Each pair of stage variables tests whether the value in that column has changed and remembers the value from the previous row (...

Because no-one can do it, let's all demand picoseconds!
After all, processors are getting faster. We may need to start talking about partial binary digits.

"It's probably all about quantum"
-- Archchancellor Mustrum Ridcully

I don't know, since TX is not my area of expertise. Certainly regular DS jobs could do it. I would imagine that you could separate processing into two streams, one for each format, and bring them together at the end (maybe by writing stream 2 into a temporary file and using an O/S command like cat t...

If you import the stored procedure definition you will have a "table definition" through which you can refer to the record structure processed by the stored procedure. After that it should be straightforward.

My point was that the answer will be different if each row is ten bytes long or if each row is ten million bytes long (yes, that's a long row, but sometimes exaggeration helps to make the point).

There is no need for concatenation operators in the date format string. "%yyyy/%mm/%dd" would suffice.

A "serious of jobs"??? I guess it is. :D

Use date_from_string (Modify stage) or StringToDate (Transformer stage) to convert the date into a string, supplying an approprate date format string. Use string_from_date (Modify stage) or DateToString (Transformer stage) to convert that result into a string of the desired format, again supplying a...

Each phantom (= background) process in DataStage directs its output to a file in the &PH& directory. The file name is made up from the program name then the time started (in DataStage internal format) then the date started (in DataStage internal format). DSD.RUN is the program that runs serv...

You could always concatenate Rnd(1000)

Error code 2 means "file not found". It's an operating system exit status.

Unless you have not posted the entire error message, it appears to be looking in a directory called "...".

There are limits on how much data the Aggregator stage can store in memory; if these are exceeded an abnormal termination occurs. For large data sets always sort incoming data on the grouping columns, and assert that they are so sorted on the Aggregator stage's input link. What this means is that, a...