[jira] [Commented] (TINKERPOP-1082) INPUT_RDD and INPUT_FORMAT are bad, we should just have one key.

["Marko A. Rodriguez (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/TINKERPOP-1082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15199707#comment-15199707 ] 

Marko A. Rodriguez commented on TINKERPOP-1082:
-----------------------------------------------

I went with {{gremlin.hadoop.graphReader}} and {{gremlin.hadoop.graphWriter}} as these terms align with the I/O package of graph readers and writers.

> INPUT_RDD and INPUT_FORMAT are bad, we should just have one key.
> ----------------------------------------------------------------
>
>                 Key: TINKERPOP-1082
>                 URL: https://issues.apache.org/jira/browse/TINKERPOP-1082
>             Project: TinkerPop
>          Issue Type: Improvement
>          Components: hadoop
>    Affects Versions: 3.1.0-incubating
>            Reporter: Marko A. Rodriguez
>             Fix For: 3.2.0-incubating
>
>
> Right now we have two keys for input to a {{HadoopGraph}}.
> * gremlin.hadoop.graphInputFormat
> * gremlin.spark.graphInputRDD
> Likewise for output. I have so many if/else checks because both of these can be set that I think we should make one and only one input.
> * gremlin.hadoop.graphInputClass
> Likewise, for output: gremlin.hadoop.graphOutputClass.
> This will make things so much simpler. However, it will break current implementations (and not backwards compatible).



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