[ 
https://issues.apache.org/jira/browse/ATLAS-1713?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Madhan Neethiraj resolved ATLAS-1713.
-------------------------------------
    Resolution: Invalid

[~ssainath] - when a type is has references from other types, it will not be 
possible to delete the type before deleting the types that reference this type. 
And it is possible for the types involved to have reference to each other - 
like hive_table & hive_column. And API that deletes a single type will not be a 
handle such cases. Hence the API that takes AtlasTypesDef, where all types to 
be deleted can be specified.

[~davidrad]
bq. I suggest that we should only allow a delete of a type that has no 
associated active instances.
bq. if there are relationships to / from other types - then potentially those 
other types are invalidated by the deletion of this type.
These were implemented in ATLAS-1673.

bq. I assume we will soft delete a type - so it is still present in the graph.
Type deletes are allowed only when no entity or other type references the type. 
The request for type delete is usually to get rid of classifications created by 
users (like testTag, testing, ..). Given this, do you think soft-delete of 
types is a high priority/critical item?

> V2 API to delete a type requires type definition to given as body 
> ------------------------------------------------------------------
>
>                 Key: ATLAS-1713
>                 URL: https://issues.apache.org/jira/browse/ATLAS-1713
>             Project: Atlas
>          Issue Type: Improvement
>          Components:  atlas-core
>    Affects Versions: 0.9-incubating
>            Reporter: Sharmadha Sainath
>            Priority: Minor
>
> V2 API 
> {code}
> api/atlas/v2/types/typedefs
> {code}
> requires body data when doing a HTTP DELETE request . 
> Since type has both name and guid , it could be given as path instead of 
> sending the whole type definition as body.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to