ASF GitHub Bot commented on BAHIR-102:

Github user emlaver commented on a diff in the pull request:

    --- Diff: sql-cloudant/README.md ---
    @@ -62,6 +62,8 @@ cloudant.protocol|https|protocol to use to transfer data: 
http or https
     cloudant.host||cloudant host url
     cloudant.username||cloudant userid
     cloudant.password||cloudant password
    +cloudant.useQuery|false|When enabled, for query not using index or view, 
_find will be used instead of _all_docs, some query predicates will be driven 
into datastore. However, RDD partition is ONE during _find, so parallel loading 
is not achieved
    --- End diff --
    Ok - I'm still having a hard time understanding the statement `some of the 
query predicates will be driven into datastore`.  What do you mean by driven? 
Does it have to do with `partitions used to drive JsonStoreRDD loading query 
result in parallel` ?
    Also, `As RDD partition is ONE during _find` - one what? set to one 
partition for the `jsonstore.rdd.partitions` configuration? 
    And for `parallel loading is not achieved` - This sounds awkward to me.  Is 
it not achieved because the `_find` endpoint only supports one partition? 

> Support pushing down query predicates using Cloudant Query / CouchDB Mango 
> Query
> --------------------------------------------------------------------------------
>                 Key: BAHIR-102
>                 URL: https://issues.apache.org/jira/browse/BAHIR-102
>             Project: Bahir
>          Issue Type: Improvement
>            Reporter: Yang Lei
>            Assignee: Yang Lei
> The feature of Cloudant Query / ChoudDB Mango Query is stated here: 
> https://blog.couchdb.org/2016/08/03/feature-mango-query/.
> Will enable pushing down first level AND column filtering to _find selector 
> to replace _all_docs for the database query; will continue using skip+limit 
> to offer partition and parallel loading.

This message was sent by Atlassian JIRA

Reply via email to