[
https://issues.apache.org/jira/browse/SPARK-15197?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15275016#comment-15275016
]
Nicholas Tietz commented on SPARK-15197:
----------------------------------------
I have started working on this and have an outstanding WIP pull request open
for it. This ticket was opened as the scope of my PR expanded based on
discussion on the PR.
> Improve documentation for countApprox and related functions
> -----------------------------------------------------------
>
> Key: SPARK-15197
> URL: https://issues.apache.org/jira/browse/SPARK-15197
> Project: Spark
> Issue Type: Documentation
> Reporter: Nicholas Tietz
>
> Within RDD, JavaRDDLike, PairRDDFunctions, and JavaPairRDD, countApprox and
> countByValueApprox are missing Scaladoc on their input parameters and return
> values; in particular, it is unclear what unit the timeout is in and what the
> meaning of the confidence value is.
> We should document these parameters clearly. This includes:
> * documenting the timeout
> * documenting/clarifying the confidence meaning/constraints
> * adding a check in the code to ensure confidence is in the required range
> * clarifying documentation in related methods (such as more detail on
> relativeSD in similar approx methods).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
