[
https://issues.apache.org/jira/browse/DRILL-7268?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Igor Guzenko updated DRILL-7268:
--------------------------------
Description:
When Hive stores array data in parquet format, it creates schema for such
columns, like:
arr_n_0 ARRAY<INT>
{code:java}
optional group arr_n_0 (LIST) {
repeated group bag {
optional int32 array_element;
}
}
{code}
Sample result before the changes was:
{code:java}
{"bag":[{"array_element":1},\{"array_element":2}]}
{code}
After the changes Drill reads only array elements data without additional keys
like "bag" or "array_element":
{code}[1,2] \{code} .
Please read Design Doc linked to parent task for more details.
was:
When Hive stores array data in parquet format, it creates schema for such
columns, like:
arr_n_0 ARRAY<INT>
{code}
optional group arr_n_0 (LIST) {
repeated group bag {
optional int32 array_element;
}
}
{code}
Sample result before the changes was:
{code}\{"bag":[{"array_element":1},\{"array_element":2}]} \{code}
After the changes Drill reads only array elements data without additional keys
like "bag" or "array_element":
{code} [1,2] \{code} .
Please read Design Doc linked to parent task for more details.
> Read Hive array with parquet native reader
> ------------------------------------------
>
> Key: DRILL-7268
> URL: https://issues.apache.org/jira/browse/DRILL-7268
> Project: Apache Drill
> Issue Type: Sub-task
> Reporter: Igor Guzenko
> Assignee: Igor Guzenko
> Priority: Major
> Labels: ready-to-commit
> Fix For: 1.17.0
>
>
> When Hive stores array data in parquet format, it creates schema for such
> columns, like:
> arr_n_0 ARRAY<INT>
> {code:java}
> optional group arr_n_0 (LIST) {
> repeated group bag {
> optional int32 array_element;
> }
> }
> {code}
> Sample result before the changes was:
> {code:java}
> {"bag":[{"array_element":1},\{"array_element":2}]}
> {code}
> After the changes Drill reads only array elements data without additional
> keys like "bag" or "array_element":
> {code}[1,2] \{code} .
>
> Please read Design Doc linked to parent task for more details.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
