DSXchange

Dears,

I want to retrieve multi-rows from a hash file to be used as reference to lookup transformer. Reference link have retrieved multi-rows when using ODBC stage with a direct reference but I need to hash data first before lookup.

I found the following reply by 'Ray' while searching:
viewtopic.php?t=88934

ray.wurlod wrote:How you create multi-valued columns depends on how you create the hashed file.
If you're using a Hashed File or UV stage to create the hashed file, specify M in the Type column in the Columns grid.

I've tried to set 'M' in the Type column for key columns as suggested but it still didn't work.

Regards.

You can not do it with a Hashed File stage. You can, however, do it with a UniVerse stage that references the hashed file through a VOC pointer.

The 'M' didn't work how? Making something 'multi-value' has absolutely nothing to do with a multi-row result set, by the way, so I guess that makes my question rhetorical.

As noted in many other posts here, you would need to build the hashed file with X fields as the "key" and then use "less than X" in the lookup when you create the VOC pointer over the hashed file and then reference it using a UV stage. Keeping the keys the same would result in still having a singleton lookup. And don't forget the CREATE.INDEX and BUILD.INDEX steps! Without that, it will be slow as, well... mole asses.

It can be done through the standard utility routine 'UtilityHashLookup' under sdk option.

Sorry, but no it cannot.

Read your Hashed File in a UV stage and look it up... Multi-row lookup will not be enabled for a HF lookup..