DSXchange

Not possible. The reject link gets rows that do not match the defined metadata. You get the input row as a raw string, in one column. This is not negotiable. There is a workaround, which is to read the file as a single VarChar column and parse it yourself in a Transformer stage. The Sequential File ...

What you see in the Director is the last reported status of the job. If the job was killed with no grace time in which to update its reported status, its status will continue to show as "Running" forever. Or until the status file is cleared. But, if you always stop job processes gracefully (never us...

Do let us know how you get on.

And also, if they continue to refuse, how they suggest you work around the fact that the Oracle Enterprise stage can not be used.

It is possible, but only using a UV stage, as noted. Performance will be helped by indexing the four columns that make up the partial key. The UV stage does support multiple reference row return (find out how by searching the forum or on-line Help).

Etiquette Note We don't do "urgent" here. DSXchange is an all-volunteer site - no-one gets paid for posting. If you want urgent help, sign up with your support provider for premium service and learn the true cost of urgent. The short answer is to count the number of items, say N, then get the Floor...

DSXchange is once again offering the Server to Parallel Transition lab, in the last week of March, in either Chicago or Dallas (to be determined). Please register your interest, and likelihood of attending, by following the link in the advertisement on the DSXchange home page (and, ideally, followin...

ODBC (BCI) functions can be called from before/after subroutines.

However, you will find that the Data Direct ODBC drivers that ship with DataStage are not licensed for this, so you would need to purchase valid licensing for the ODBC drivers. This is not cheap.

Is the receiving field big enough to receive the data?

Not when using the system tables is "hard-wired" into the stage! Presumably because this is the most reliable way.
In any case, it's documented as being required.

Craig wrote:same company in different departments and sit on different floors

Can't be much of a workplace if you have to sit on the floor!

My guess is that you need the Transformer stage downstream of an Aggregator stage.

And, of course, data pre-sorted on the aggregation (grouping) columns.

Unless, of course, the delimiter character appears in data - in that case you DO need quoted character strings.

Technically the double quotes are only required when an identifier (table name, column name, etc.) is a reserved word or in the vocabulary (VOC) file. It's always safe to use them, however. TYPE is a reserved word in UniVerse/SQL, it is used in CREATE TABLE statement to specify the hashing algorithm...

The buffer size is specified in KBytes, not in bytes.

Yes you can. What error message do you receive from UniVerse? I just ran a similar query (the only difference being the column names) in UniVerse. SELECT "TYPE","F2" FROM VOC GROUP BY "TYPE","F2" HAVING COUNT(*) > 1; I also checked with a space between...