Search found 53125 matches

by ray.wurlod
Mon Jul 18, 2005 2:52 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Sort VS. AGGR
Replies: 4
Views: 1043

Do you need to sort the entire DataSet? Or just the data on each partition? In the latter case, the PX sort may well be faster.
by ray.wurlod
Mon Jul 18, 2005 2:50 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Error in Aggregator
Replies: 2
Views: 927

In SQL terminology it's complaining about an attempt to

Code: Select all

SELECT BI_IM_TP_SYSTEM_ID, SUM(BI_IM_TP_SYSTEM_ID) FROM inputlink GROUP BY BI_IM_TP_SYSTEM_ID
by ray.wurlod
Mon Jul 18, 2005 2:47 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Working with Fixed width files
Replies: 4
Views: 1135

If your file really is fixed width it will look like vin shelke You need to be absolutely sure not only that the file really is fixed width but also that you have the column widths perfectly correct. If the file is actually some delimited format, then you need metadata to describe that format correc...
by ray.wurlod
Mon Jul 18, 2005 2:44 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: CSV file to Oracle
Replies: 4
Views: 1442

There's a file called ora.11858.910475.0.log.bad that will tell you what the problem in sqlldr was. You can use the oerr command to decode the Oracle error numbers, and usually get advice about how to remedy the errors.
by ray.wurlod
Mon Jul 18, 2005 2:42 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: Frequency of a record in a sequential file
Replies: 4
Views: 1234

Currently there are no manuals available for download.

This may change now that IBM has acquired the products, as their model includes publishing manuals on a web site.

But not yet.
by ray.wurlod
Mon Jul 18, 2005 2:39 pm
Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
Topic: More info from SeqFileStg (Input) Reject Link
Replies: 3
Views: 989

You will note that you can not influence the metadata on this link, therefore the answer is NO as far as the Sequential File stage is concerned. The rows on this link simply don't match the metadata defined on the output link's schema. Possibly you could direct that reject link into some other stage...
by ray.wurlod
Mon Jul 18, 2005 1:20 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Multiple input tables and one target table
Replies: 11
Views: 2573

If you have prepared SQL then clearly it can be done with user-defined SQL. Yes it's a horrible query but hopefully the query optimizer can deal with it sensibly. Some bitmap indexes, if these are possible, could make such a query really fly.
by ray.wurlod
Mon Jul 18, 2005 1:04 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Multiple input tables and one target table
Replies: 11
Views: 2573

Of course. User-defined SQL in a regular stage type. 8)

Do "insert or update" into the target table. Normally I advise against this but, with the complex selection requirements and the small volumes, you'll get away with it.
by ray.wurlod
Mon Jul 18, 2005 1:02 am
Forum: Data Integration
Topic: Good grief, have you seen Ascential World's new name?
Replies: 5
Views: 4325

So, if session details aren't available, could one be buying a pig in a poke?
:?:
by ray.wurlod
Mon Jul 18, 2005 12:28 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: wait for file activity
Replies: 2
Views: 970

Is there any DOS command to check the size of the file greater than 0kb and to delete the file? If you're on 7.5 with MKS toolkit, or have MKS toolkit anyway, then you can write a UNIX script to do this. Something like: if [ -z $filename ] then rm $filename fi I don't know if there's a DOS equivale...
by ray.wurlod
Mon Jul 18, 2005 12:23 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Replacing Carriage Returns in fields
Replies: 9
Views: 3611

Be careful if you're really on Windows, which your original post specifies. The Windows line termination sequence is actually two characters, made up of a line feed and a carriage return. Since Convert() only does character-by-character conversion, you would definitely need Ereplace() - which does s...
by ray.wurlod
Mon Jul 18, 2005 12:20 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: calculation for new record based on multiple existingrecords
Replies: 3
Views: 750

You need joining (perhaps self-joining), grouping and possibly set functions when extracting from SQL Server.
by ray.wurlod
Mon Jul 18, 2005 12:18 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Multiple input tables and one target table
Replies: 11
Views: 2573

Do it all in DataStage and ditch the PL/SQL? Each time you invoke sqlplus there's all the overhead of command line processing and parsing, waking an Oracle listener, getting the SQL organized (optimized?), and so on. DataStage establishes one connection, has the SQL prepared once, and then only pass...
by ray.wurlod
Mon Jul 18, 2005 12:16 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: From Hash file to loading Table
Replies: 9
Views: 1674

You could use the Debugger if they hadn't broken it. :cry:
by ray.wurlod
Mon Jul 18, 2005 12:13 am
Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
Topic: Import error
Replies: 5
Views: 1284

Almost certainly you (the executing user) lack permission to delete these dynamic hashed files. (Among other things, you need write permission to the parent directory, as well as write permission to the hashed file directories themselves.)