[
https://issues.apache.org/jira/browse/HUDI-8898?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Y Ethan Guo updated HUDI-8898:
------------------------------
Description:
The INSERT SQL statement with a subset of columns fails on Hudi table in Spark
3.5; the same succeeds in Spark 3.4.
Create table:
{code:java}
spark.sql(
s"""
|create table $tableName (
| id int,
| dt string,
| name string,
| price double,
| ts long
|) using hudi
| tblproperties (primaryKey = 'id')
| $createTablePartitionClause
| location '$tablePath'
""".stripMargin) {code}
INSERT INTO with a subset of columns
{code:java}
spark.sql(
s"""
| insert into $tableName (dt, ts, name, id)
| values ('2025-01-04', 4000, 'a4', 4)
""".stripMargin){code}
It fails with
{code:java}
[INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
`spark_catalog`.`default`.`h1`, the reason is not enough data columns:
Table columns: `id`, `name`, `price`, `ts`, `dt`.
Data columns: `dt`, `ts`, `name`, `id`.
org.apache.spark.sql.AnalysisException:
[INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
`spark_catalog`.`default`.`h1`, the reason is not enough data columns:
Table columns: `id`, `name`, `price`, `ts`, `dt`.
Data columns: `dt`, `ts`, `name`, `id`.
at
org.apache.spark.sql.errors.QueryCompilationErrors$.cannotWriteNotEnoughColumnsToTableError(QueryCompilationErrors.scala:2126)
at
org.apache.spark.sql.catalyst.analysis.TableOutputResolver$.resolveOutputColumns(TableOutputResolver.scala:70)
at
org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns(HoodieSpark3CatalystPlanUtils.scala:51)
at
org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns$(HoodieSpark3CatalystPlanUtils.scala:46)
at
org.apache.spark.sql.HoodieSpark35CatalystPlanUtils$.resolveOutputColumns(HoodieSpark35CatalystPlanUtils.scala:32)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.coerceQueryOutputColumns(InsertIntoHoodieTableCommand.scala:168)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.alignQueryOutput(InsertIntoHoodieTableCommand.scala:145)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.run(InsertIntoHoodieTableCommand.scala:99)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand.run(InsertIntoHoodieTableCommand.scala:62)
{code}
was:
The
Create table:
{code:java}
spark.sql(
s"""
|create table $tableName (
| id int,
| dt string,
| name string,
| price double,
| ts long
|) using $format
| tblproperties (primaryKey = 'id')
| $createTablePartitionClause
| location '$tablePath'
""".stripMargin) {code}
INSERT INTO with a subset of columns
{code:java}
spark.sql(
s"""
| insert into $tableName (dt, ts, name, id)
| values ('2025-01-04', 4000, 'a4', 4)
""".stripMargin){code}
It fails with
{code:java}
[INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
`spark_catalog`.`default`.`h1`, the reason is not enough data columns:
Table columns: `id`, `name`, `price`, `ts`, `dt`.
Data columns: `dt`, `ts`, `name`, `id`.
org.apache.spark.sql.AnalysisException:
[INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
`spark_catalog`.`default`.`h1`, the reason is not enough data columns:
Table columns: `id`, `name`, `price`, `ts`, `dt`.
Data columns: `dt`, `ts`, `name`, `id`.
at
org.apache.spark.sql.errors.QueryCompilationErrors$.cannotWriteNotEnoughColumnsToTableError(QueryCompilationErrors.scala:2126)
at
org.apache.spark.sql.catalyst.analysis.TableOutputResolver$.resolveOutputColumns(TableOutputResolver.scala:70)
at
org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns(HoodieSpark3CatalystPlanUtils.scala:51)
at
org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns$(HoodieSpark3CatalystPlanUtils.scala:46)
at
org.apache.spark.sql.HoodieSpark35CatalystPlanUtils$.resolveOutputColumns(HoodieSpark35CatalystPlanUtils.scala:32)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.coerceQueryOutputColumns(InsertIntoHoodieTableCommand.scala:168)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.alignQueryOutput(InsertIntoHoodieTableCommand.scala:145)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.run(InsertIntoHoodieTableCommand.scala:99)
at
org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand.run(InsertIntoHoodieTableCommand.scala:62)
{code}
> Support INSERT SQL statement with a subset of columns in Spark 3.5
> ------------------------------------------------------------------
>
> Key: HUDI-8898
> URL: https://issues.apache.org/jira/browse/HUDI-8898
> Project: Apache Hudi
> Issue Type: Sub-task
> Reporter: Y Ethan Guo
> Assignee: Y Ethan Guo
> Priority: Blocker
> Fix For: 1.0.1
>
>
> The INSERT SQL statement with a subset of columns fails on Hudi table in
> Spark 3.5; the same succeeds in Spark 3.4.
> Create table:
> {code:java}
> spark.sql(
> s"""
> |create table $tableName (
> | id int,
> | dt string,
> | name string,
> | price double,
> | ts long
> |) using hudi
> | tblproperties (primaryKey = 'id')
> | $createTablePartitionClause
> | location '$tablePath'
> """.stripMargin) {code}
> INSERT INTO with a subset of columns
> {code:java}
> spark.sql(
> s"""
> | insert into $tableName (dt, ts, name, id)
> | values ('2025-01-04', 4000, 'a4', 4)
> """.stripMargin){code}
> It fails with
> {code:java}
> [INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
> `spark_catalog`.`default`.`h1`, the reason is not enough data columns:
> Table columns: `id`, `name`, `price`, `ts`, `dt`.
> Data columns: `dt`, `ts`, `name`, `id`.
> org.apache.spark.sql.AnalysisException:
> [INSERT_COLUMN_ARITY_MISMATCH.NOT_ENOUGH_DATA_COLUMNS] Cannot write to
> `spark_catalog`.`default`.`h1`, the reason is not enough data columns:
> Table columns: `id`, `name`, `price`, `ts`, `dt`.
> Data columns: `dt`, `ts`, `name`, `id`.
> at
> org.apache.spark.sql.errors.QueryCompilationErrors$.cannotWriteNotEnoughColumnsToTableError(QueryCompilationErrors.scala:2126)
> at
> org.apache.spark.sql.catalyst.analysis.TableOutputResolver$.resolveOutputColumns(TableOutputResolver.scala:70)
> at
> org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns(HoodieSpark3CatalystPlanUtils.scala:51)
> at
> org.apache.spark.sql.HoodieSpark3CatalystPlanUtils.resolveOutputColumns$(HoodieSpark3CatalystPlanUtils.scala:46)
> at
> org.apache.spark.sql.HoodieSpark35CatalystPlanUtils$.resolveOutputColumns(HoodieSpark35CatalystPlanUtils.scala:32)
> at
> org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.coerceQueryOutputColumns(InsertIntoHoodieTableCommand.scala:168)
> at
> org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.alignQueryOutput(InsertIntoHoodieTableCommand.scala:145)
> at
> org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand$.run(InsertIntoHoodieTableCommand.scala:99)
> at
> org.apache.spark.sql.hudi.command.InsertIntoHoodieTableCommand.run(InsertIntoHoodieTableCommand.scala:62)
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
