[
https://issues.apache.org/jira/browse/PARQUET-1881?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17153344#comment-17153344
]
Gabor Szadovszky commented on PARQUET-1881:
-------------------------------------------
Thanks for the reference.
I did not find anything related to this in the {{parquet-mr}} code base. Even
though this object is added to the thrift file 7 years ago it seems it is never
written or used in {{parquet-mr}}. Meanwhile, it seems the Impala
implementation writes this objects and probably uses it as well. I am not an
Impala expert, though.
> How to enable sorted array flag while writing a column using parquet-mr
> -----------------------------------------------------------------------
>
> Key: PARQUET-1881
> URL: https://issues.apache.org/jira/browse/PARQUET-1881
> Project: Parquet
> Issue Type: Task
> Components: parquet-mr
> Reporter: Khasim Shaik
> Priority: Blocker
> Labels: newbie
>
> I want to understand how can we enable the flag "sortedArray" information in
> metadata while writing a row group or column
> I am exploring parquet.thrift to understand more about metadata,
> I observed a field in metadata which is related to below struct in
> parquet.thrift
> I am wondering how to set these fields from parquet while writing a column or
> rowgroup
> struct SortingColumn {
> /** The column index (in this row group) **/
> 1: required i32 column_idx
> /** If true, indicates this column is sorted in descending order. **/
> 2: required bool descending
> /** If true, nulls will come before non-null values, otherwise,
> * nulls go at the end. */
> 3: required bool nulls_first
> }
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
