[ 
https://issues.apache.org/jira/browse/COUCHDB-2651?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14393492#comment-14393492
 ] 

ASF subversion and git services commented on COUCHDB-2651:
----------------------------------------------------------

Commit 985c952aa2274664107472675fe37eea105085b3 in couchdb-mango's branch 
refs/heads/2651-delete-bulk-docs from [~tonysun83]
[ https://git-wip-us.apache.org/repos/asf?p=couchdb-mango.git;h=985c952 ]

Return success and failure information

Rather than just return "indexes deleted", we return a list of successful 
deletes
and failed deletes along with the ids. For a single delete, we still throw the 
same
errors as before to be backwards compatible with any applications expecting an 
error.

Fixes COUCHDB-2651


> Deleting a Mango Design Docs using bulk docs
> --------------------------------------------
>
>                 Key: COUCHDB-2651
>                 URL: https://issues.apache.org/jira/browse/COUCHDB-2651
>             Project: CouchDB
>          Issue Type: Bug
>      Security Level: public(Regular issues) 
>          Components: go
>            Reporter: Robert Kowalski
>
> In Fauxtont the user can list documents / indexes and select multiple of 
> them. The user can then press one button to delete all selected documents. 
> This works for Mango created design docs, but it is quite hard to get the 
> revision id as /_index does not list them. Current workarounds are:
> 1. firing a lot of multiple requests, but given that one doc list can have 
> 100 entries this is not a viable solution:
> curl -XGET 
> "http://localhost:8000/animaldb/_design/e4d338e5d6f047749f5399ab998b4fa04ba0c816";
> {"_id":"_design/e4d338e5d6f047749f5399ab998b4fa04ba0c816","_rev":"1-1f37105891681e717ae37dae973bc6a4","language":"query","views":{"e4d338e5d6f047749f5399ab998b4fa04ba0c816":{"map":{"fields":{"_id":"asc"}},"reduce":"_count","options":{"def":{"fields":["_id"]},"w":2}}}}
> 2. Querying all the design docs with includedocs, and don't use the _index 
> endpoint, then filter out all all indexes which are not mango indexes and use 
> the result for building the list of documents which then can get checked and 
> deleted (downside: performance, not using the existing endpoint from mango 
> which makes it a bit obsolete, at least for Fauxton)
> For me there are two possible solutions:
>  - returning the rev id in GET /_index
>  - implementing an additional POST /_bulk_index endpoint
> while I would prefer the first as the _bulk_docs endpoint already exists and 
> works very well.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to