Search found 4992 matches
- Mon Feb 07, 2005 3:48 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Lookup problem with a multi-instance job - Strange !
- Replies: 30
- Views: 4714
The security column is part of 3 columns that make up a key. If you were to describe to me a situation where a row is found/notfound when one of those 3 columns has other rows like itself, my guess would be 99% right that you have a metadata issue. What you write to a hash file has to be read from t...
- Mon Feb 07, 2005 3:40 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: two tables with single ODBC
- Replies: 3
- Views: 1263
- Mon Feb 07, 2005 3:38 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Lookup problem with a multi-instance job - Strange !
- Replies: 30
- Views: 4714
When I say one security I mean a set of rows with the same value under the column security. Then those rows mean nothing to each other, because they are different natural key combinations. Therefore, your problem is most likely NOT related to this column, unless it just happens to be pointing to an...
- Mon Feb 07, 2005 2:45 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Lookup problem with a multi-instance job - Strange !
- Replies: 30
- Views: 4714
Well, Kim and I verified your example data should work. Now how about actual data? From your stated example, "security" is only 1 column of 3 on which you have to match a lookup. When you say "one security in the input file", are you saying one row, or one set of rows with the same "security" value,...
- Mon Feb 07, 2005 2:17 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Lookup problem with a multi-instance job - Strange !
- Replies: 30
- Views: 4714
Input data: SECURITY START DATE TYPE RATE 101 1/1/2001 00:00:00 MATURITY 100 Now here is my lookup hash file - SECURITY START DATE TYPE SURROGATEKEY 101 1/1/2001 00:00:00 MATURITY 111 In this case, the input row should match the hash lookup based on what you are saying. If the lookup row is already...
- Mon Feb 07, 2005 1:15 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Lookup problem with a multi-instance job - Strange !
- Replies: 30
- Views: 4714
If you split your data into 8 parts, and run 8 instances to process the data, you need to make sure that rows that need to be handled together go to the same instance and you DON'T use any caching. Think of this as a single job, whereby the first row has to be seen by the second row. You can only do...
- Sat Feb 05, 2005 12:18 am
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Connection Problem AIX Server
- Replies: 8
- Views: 1943
Just to be absolutely sure, the jobs are running fine, it's the GUI clients that are on your PC that are having problems. Not just Director, but Manager and Designer as well. Are you running a lot of jobs when the clients lose their connections? A heavily loaded server with a lot of activity can cau...
- Sat Feb 05, 2005 12:06 am
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Connection Problem AIX Server
- Replies: 8
- Views: 1943
- Fri Feb 04, 2005 7:37 pm
- Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
- Topic: How to check how many clients are connected to the project?
- Replies: 6
- Views: 1935
Actually, being connected will not prevent an export; it's only when they have objects open for editing that that object is skipped by the export process. It is sufficient, therefore, to check for locked design objects (though safer not to have any user connected). You can determine both using the ...
- Fri Feb 04, 2005 3:31 pm
- Forum: IBM<sup>®</sup> DataStage Enterprise Edition (Formerly Parallel Extender/PX)
- Topic: How to check how many clients are connected to the project?
- Replies: 6
- Views: 1935
- Fri Feb 04, 2005 12:53 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: converting rows to columns
- Replies: 6
- Views: 1661
Aggregator is easiest, stage variables can be lengthy and you have that pesky problem of the last row to do with. Setup a transformer that has an output link to an aggregator. Your output link has four columns. For each row, map the source column to the appropriate target column. If that row is not ...
- Fri Feb 04, 2005 8:37 am
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Testing with null arguements
- Replies: 6
- Views: 2167
- Fri Feb 04, 2005 8:24 am
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Testing with null arguements
- Replies: 6
- Views: 2167
- Fri Feb 04, 2005 6:44 am
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: Testing with null arguements
- Replies: 6
- Views: 2167
There's no way in the test box to enter a NULL value. Put a line of code into the routine to switch the passed argument for null, such as:
Code: Select all
If Arg1 = "SWITCH THIS TO NULL FOR TESTING PURPOSES" Then
InputValue = @NULL
Else
InputValue = Arg1
End- Thu Feb 03, 2005 8:36 pm
- Forum: IBM<sup>®</sup> Infosphere DataStage Server Edition
- Topic: java calling a datastage job
- Replies: 5
- Views: 2197