yurmix opened a new issue #6519: Lookups should be able to return list of values URL: https://github.com/apache/incubator-druid/issues/6519 Hi, This is an early discussion before I proceed with a clear proposal. **The requirement:** We have a requirement in which a Registered lookup extraction function (`registeredLookup`) will return multi-value list instead of String. A common example would be tag heirarchy, in which the lookup input is the heirarchy root and the output is a flattened list for the heirarchy. Since this is not a slowly changing dimension (Which is handled at ingestion time), it requires hadeling via query time lookups. @sandeep217 described the requirement pretty well in the below issue from 2.5 years ago. Thought it would be better to open a fresh one: **[Support for lookup into a multi-value dimension #2374](https://github.com/apache/incubator-druid/issues/2374)** Lookups currently support a 1x1 and Nx1 mapping. Often there's a need to do a lookup against 1xN mapping. A common example is a lookup that maps userid to user-groups where a user may belong to more than one user groups. Denormalizing at ETL may not be desirable if user to user-group mapping changes over time. **Implementation** I'm looking for a bit of help with the design. I'm aware that this will require support in both the filtering part of the query as well as the grouping part. I was looking into adding a companion method to `LookupExtractor#apply()` that returns `Map<String, Object>` and add a call to it in the places that call `apply()` is called today. (I thought it would be better not to change apply() itself for the sake of existing written lookup extensions). @b-slim @drcrallen I saw that both of you worked on releated items in the past, perhaps you care to comment?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
