[
https://issues.apache.org/jira/browse/PARQUET-435?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15107511#comment-15107511
]
Wes McKinney commented on PARQUET-435:
--------------------------------------
I haven't started working on anything yet; I mainly need PARQUET-434 right this
second.
I was thinking about having each column reader able to populate a preallocated
C array of an indicated size from a row group (returning the actual length if
there is not enough data available in a row group), as opposed to having one
function call per value.
Aside: the lack of unit tests is causing me some anxiety -- I will put up a
patch for PARQUET-437 so that we can start asking for unit tests with further
libparquet patches.
> Provide vectorized ColumnReader interface
> -----------------------------------------
>
> Key: PARQUET-435
> URL: https://issues.apache.org/jira/browse/PARQUET-435
> Project: Parquet
> Issue Type: New Feature
> Components: parquet-cpp
> Reporter: Wes McKinney
>
> Related to PARQUET-433. The library user should be able to retrieve a batch
> of column values, repetition levels, or definition levels with a particular
> size into a preallocated C array.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
