[jira] [Commented] (OAK-1952) [Query] There should be a way to sort results from query according number of properties/predicates match

["Thomas Mueller (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OAK-1952?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14053563#comment-14053563 ] 

Thomas Mueller commented on OAK-1952:
-------------------------------------

Could you provide some information about the data?:

* Is using an asynchronous index good enough (results might not reflect the very latest changes)?
* Is propertyA is a multi-valued property? 
* What data type does it have, or do you need support for multiple data types?
* What are typical values (example values)? 
* How many nodes are there in the repository with that property?
* How many values does it typically have per node (min, max, average)? 
* How many distinct values are there? 
* Is the list of values fixed (relatively) fixed or does it change a lot? 

(I have an idea on how to implement this: using a bitmap index; but it's a bit unusual to use such an index in this area).

> [Query] There should be a way to sort results from query according number of properties/predicates match
> --------------------------------------------------------------------------------------------------------
>
>                 Key: OAK-1952
>                 URL: https://issues.apache.org/jira/browse/OAK-1952
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: query
>    Affects Versions: 1.0
>            Reporter: Vikas Saurabh
>
> Use Case: Query for value of propertyA for valueA OR valueB OR valueC and sort it according to number of matches (ie a node matching all 3 should have higher score than those mathing 1 or 2 of those).
> Currently, the only way is to fetch the result from OR query in application and then post process it. While this is fine if application is interested in all the results. But, if the query wants to fetch a page of that result set (result N to result M), then the application still needs to fetch all nodes while the platform can (possibly) optimize this internally.



--
This message was sent by Atlassian JIRA
(v6.2#6252)