[
https://issues.apache.org/jira/browse/ARROW-10706?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Weiyang Zhao updated ARROW-10706:
---------------------------------
Description:
The below code will raise IndexError:
{{dataset = pq.ParquetDataset(}}
base_path, filesystem=fs,
filters=[('string', '=', "notExisted")],
use_legacy_dataset=True
{{)}}
when the partition 'string' does not have a matching partition value
'notExisted'.
The correct behavior should be returning an empty dataset with the actual
schema.
was:
The below code will raise IndexError:
{{dataset = pq.ParquetDataset(}}
base_path, filesystem=fs,
filters=[('string', '=', "notExisted")],
use_legacy_dataset=True
{{)}}
when the partition 'string' does not have a matching partition value
'notExisted'.
> [Python][Parquet] Fix a bug when partition filters has empty result
> -------------------------------------------------------------------
>
> Key: ARROW-10706
> URL: https://issues.apache.org/jira/browse/ARROW-10706
> Project: Apache Arrow
> Issue Type: Bug
> Reporter: Weiyang Zhao
> Assignee: Weiyang Zhao
> Priority: Major
>
> The below code will raise IndexError:
> {{dataset = pq.ParquetDataset(}}
> base_path, filesystem=fs,
> filters=[('string', '=', "notExisted")],
> use_legacy_dataset=True
> {{)}}
> when the partition 'string' does not have a matching partition value
> 'notExisted'.
> The correct behavior should be returning an empty dataset with the actual
> schema.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
