DSXchange

If you're consistent and systematic with the comment field on batches promoted with Version Control, you'd be surprised how much help this can be when searching for the version where something changed.

These things look so much better if you wrap them in the Code tags.

Then fiddle with Preview till it looks perfect.

Then submit.

Is the source from which you're reading, or the source from which you're loading the hashed file, a Sequential File? Also, when you fail to find the key in a reference lookup, documented DataStage behaviour is to return NULL for all columns on the reference input link; if you don't check for this, y...

I DID say roughly! Most operators and functions have equivalents for handling multi-valued data, in most cases formed by appending "S" to the single-valued name. For example LEN becomes LENS, FIELD becomes FIELDS, NE becomes NES, OCONV becomes OCONVS and so on. Single-valued constants quantities may...

Yes. Many. All have the effect of opening the SDKSequences file, locking the appropriate record for update, reading that record (optional), writing the reset value, releasing the lock, and closing the file. Optionally you can build all of that into a transaction. Is this not exactly what the SDK rou...

I like it! Let's hope that everyone reads the new category names and blurbs about what belongs where, and complies. Time for a NewsLetter to announce the new layout, and to encourage folk to choose the right forum. Maybe (at least for a time) add "formerly" to the DataStage forum titles, for example...

If you're running version 7.1 or later, there are compilation checkboxes in the Job Sequence that effect exactly this functionality. On re-run, those jobs that completed successfully last time are skipped (with a log entry indicating that they've been skipped). In earlier releases, you're up for a D...

The I-descriptor expression has roughly the following form. If Oconv(SQLDataTypeCode,"R;-7,12") Then Field("Bit|TinyInt|BigInt|LongVarBinary|VarBinary|Binary|LongVarChar|VarChar|Char|Numeric|Decimal|Integer|SmallInt|Float|Real|Double|Date|Time|TimeStamp|VarChar", &quo...

Currently, only the ODBC and UV stages support multi-row returns from reference lookups. Read the on-line help topic Defining Multirow Lookup for Reference Inputs . --> Thanks for ur input Ray.. I know that only Universe and ODBC stages return multiple rows.. but I just want to return a single valu...

Sounds like "they" have to add a chmod to the process.

There is definitly no intrinsic function for any form of interval handling.

Of course, once you've built the Routine, it's built-in for ever more! (Looks kinda like one of mine.)

As well as specifying the location of data in the hashed file, the file dictionary specifies display format. One aspect of display format is a default column heading. This does not have to be the same as the column name. That's what's happening here. Field names in file dictionaries are specified in...

Ray, "The assumption is that, in a reference lookup, you're retrieving one row, so that GROUP BY, HAVING and ORDER BY are irrelevant." --> Consider this query select max(colA),colB from table A where colB=<some value> If I have to put this query in the reference table I need a group by on colB - th...

Currently, only the ODBC and UV stages support multi-row returns from reference lookups. Read the on-line help topic Defining Multirow Lookup for Reference Inputs.

You just need a couple of extra entries in uvodbc.config (hostname and server).