Posted: Thu Jan 23, 2003 10:53 pm
I believe conditional lookups are planned for future releases of Datastage.
There are some alternatives:
- Split your data streams to let the minimum number of rows advance to each DBLookup tranformer. Those rows that are successfully lookup up are outputted to a file, the rest are passed on to the next transform:
[HashFile] [DBLookup1] [DBLookup2] |
[FlatFile]->[Transform]->[Transform]->[Transform]->[FlatFile]
------------->[FlatFile]
------------------------->[FlatFile]
Merge the three files with an after job command. This is much faster since you are not carrying the bulk of your data through your 2nd and 3rd transforms as extra baggage.
- You could try to build a hash file lookup to replace the dblookup1 and dblookup2 and populate it with every possible combination of key and computed value. This really depends on your data and lookup SQL.
regards
Vincent
There are some alternatives:
- Split your data streams to let the minimum number of rows advance to each DBLookup tranformer. Those rows that are successfully lookup up are outputted to a file, the rest are passed on to the next transform:
[HashFile] [DBLookup1] [DBLookup2] |
[FlatFile]->[Transform]->[Transform]->[Transform]->[FlatFile]
------------->[FlatFile]
------------------------->[FlatFile]
Merge the three files with an after job command. This is much faster since you are not carrying the bulk of your data through your 2nd and 3rd transforms as extra baggage.
- You could try to build a hash file lookup to replace the dblookup1 and dblookup2 and populate it with every possible combination of key and computed value. This really depends on your data and lookup SQL.
regards
Vincent