[DRILL-7934] NullPointerException error when reading parquet files

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.18.0
• Fix Version/s: 1.19.0
• Component/s: Storage - Parquet

Environment

Drill 1.18
Ambari 2.7.4
Spark 3.0.2

• External issue URL: https://issues.apache.org/jira/browse/DRILL-7769
• Assignee: Unassigned
• Reporter: cdmikechen
• Reviewer: Cong Luo
• Created: 26/May/21 01:05
• Updated: 01/Jun/21 00:24
• Resolved: 01/Jun/21 00:24

Description

I create a dataset using spark ml, when I use drill 1.18 to query this dataset folder, it report error this:

[Error Id: 92d3f331-ffca-46b5-a64c-87453b88a108 on xxx.xxx.xxx:31010]
        at org.apache.drill.common.exceptions.UserException$Builder.build(UserException.java:657)
        at org.apache.drill.exec.work.foreman.Foreman$ForemanResult.close(Foreman.java:788)
        at org.apache.drill.exec.work.foreman.QueryStateProcessor.checkCommonStates(QueryStateProcessor.java:322)
        at org.apache.drill.exec.work.foreman.QueryStateProcessor.planning(QueryStateProcessor.java:216)
        at org.apache.drill.exec.work.foreman.QueryStateProcessor.moveToState(QueryStateProcessor.java:76)
        at org.apache.drill.exec.work.foreman.Foreman.run(Foreman.java:300)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:748)
Caused by: org.apache.drill.exec.work.foreman.ForemanException: Unexpected exception during fragment initialization: Error while applying rule DrillPushProjectIntoScanRule:enumerable, args [rel#478:LogicalProject.NONE.ANY([]).[](input=RelSubset#477,label=$1), rel#452:EnumerableTableScan.ENUMERABLE.ANY([]).[](table=[hdfs_dataset.default, /home/spark/dataset/default/test2/*.parquet])]
        at org.apache.drill.exec.work.foreman.Foreman.run(Foreman.java:301)
        ... 3 common frames omitted
Caused by: java.lang.RuntimeException: Error while applying rule DrillPushProjectIntoScanRule:enumerable, args [rel#478:LogicalProject.NONE.ANY([]).[](input=RelSubset#477,label=$1), rel#452:EnumerableTableScan.ENUMERABLE.ANY([]).[](table=[hdfs_dataset.default, /home/spark/dataset/default/test2/*.parquet])]
        at org.apache.calcite.plan.volcano.VolcanoRuleCall.onMatch(VolcanoRuleCall.java:235)
        at org.apache.calcite.plan.volcano.VolcanoPlanner.findBestExp(VolcanoPlanner.java:633)
        at org.apache.calcite.tools.Programs$RuleSetProgram.run(Programs.java:327)
        at org.apache.drill.exec.planner.sql.handlers.DefaultSqlHandler.transform(DefaultSqlHandler.java:405)
        at org.apache.drill.exec.planner.sql.handlers.DefaultSqlHandler.transform(DefaultSqlHandler.java:351)
        at org.apache.drill.exec.planner.sql.handlers.DefaultSqlHandler.convertToRawDrel(DefaultSqlHandler.java:245)
        at org.apache.drill.exec.planner.sql.handlers.DefaultSqlHandler.convertToDrel(DefaultSqlHandler.java:308)
        at org.apache.drill.exec.planner.sql.handlers.DefaultSqlHandler.getPlan(DefaultSqlHandler.java:173)
        at org.apache.drill.exec.planner.sql.DrillSqlWorker.getQueryPlan(DrillSqlWorker.java:283)
        at org.apache.drill.exec.planner.sql.DrillSqlWorker.getPhysicalPlan(DrillSqlWorker.java:163)
        at org.apache.drill.exec.planner.sql.DrillSqlWorker.convertPlan(DrillSqlWorker.java:140)
        at org.apache.drill.exec.planner.sql.DrillSqlWorker.getPlan(DrillSqlWorker.java:93)
        at org.apache.drill.exec.work.foreman.Foreman.runSQL(Foreman.java:593)
        at org.apache.drill.exec.work.foreman.Foreman.run(Foreman.java:274)
        ... 3 common frames omitted
Caused by: java.lang.NullPointerException: null
        at org.apache.drill.exec.store.parquet.ParquetGroupScanStatistics.checkForPartitionColumn(ParquetGroupScanStatistics.java:186)
        at org.apache.drill.exec.store.parquet.ParquetGroupScanStatistics.collect(ParquetGroupScanStatistics.java:119)
        at org.apache.drill.exec.store.parquet.ParquetGroupScanStatistics.<init>(ParquetGroupScanStatistics.java:59)
        at org.apache.drill.exec.store.parquet.BaseParquetMetadataProvider.getParquetGroupScanStatistics(BaseParquetMetadataProvider.java:293)
        at org.apache.drill.exec.store.parquet.BaseParquetMetadataProvider.getTableMetadata(BaseParquetMetadataProvider.java:249)
        at org.apache.drill.exec.store.parquet.BaseParquetMetadataProvider.initializeMetadata(BaseParquetMetadataProvider.java:203)
        at org.apache.drill.exec.store.parquet.BaseParquetMetadataProvider.init(BaseParquetMetadataProvider.java:170)
        at org.apache.drill.exec.metastore.store.parquet.ParquetTableMetadataProviderImpl.<init>(ParquetTableMetadataProviderImpl.java:95)
        at org.apache.drill.exec.metastore.store.parquet.ParquetTableMetadataProviderImpl.<init>(ParquetTableMetadataProviderImpl.java:48)
        at org.apache.drill.exec.metastore.store.parquet.ParquetTableMetadataProviderImpl$Builder.build(ParquetTableMetadataProviderImpl.java:415)
        at org.apache.drill.exec.store.parquet.ParquetGroupScan.<init>(ParquetGroupScan.java:150)
        at org.apache.drill.exec.store.parquet.ParquetGroupScan.<init>(ParquetGroupScan.java:120)
        at org.apache.drill.exec.store.parquet.ParquetFormatPlugin.getGroupScan(ParquetFormatPlugin.java:202)
        at org.apache.drill.exec.store.parquet.ParquetFormatPlugin.getGroupScan(ParquetFormatPlugin.java:79)
        at org.apache.drill.exec.store.dfs.FileSystemPlugin.getPhysicalScan(FileSystemPlugin.java:226)
        at org.apache.drill.exec.store.dfs.FileSystemPlugin.getPhysicalScan(FileSystemPlugin.java:209)
        at org.apache.drill.exec.planner.logical.DrillTable.getGroupScan(DrillTable.java:119)
        at org.apache.drill.exec.planner.logical.DrillPushProjectIntoScanRule.canPushProjectIntoScan(DrillPushProjectIntoScanRule.java:190)
        at org.apache.drill.exec.planner.logical.DrillPushProjectIntoScanRule.onMatch(DrillPushProjectIntoScanRule.java:107)
        at org.apache.calcite.plan.volcano.VolcanoRuleCall.onMatch(VolcanoRuleCall.java:208)
        ... 16 common frames omitted

It is same like issue https://issues.apache.org/jira/browse/DRILL-7769.
I add some log information and found this:

TRACE o.a.d.e.s.p.ParquetGroupScanStatistics - check schema path `features`.`values`.`list`.`element` with major type null
 current partitionColTypeMap = {`features`.`indices`.`list`.`element`=null, `features`.`type`=minor_type: TINYINT
mode: REQUIRED
, `features`.`size`=minor_type: INT
mode: OPTIONAL
}
2021-05-25 15:39:21,066 [1f535658-f840-9f0e-1a7b-21080514bb7b:foreman] TRACE o.a.d.e.s.p.ParquetGroupScanStatistics - check schema path `label` with major type minor_type: FLOAT8
mode: REQUIRED

 current partitionColTypeMap = {`features`.`indices`.`list`.`element`=null, `features`.`type`=minor_type: TINYINT
mode: REQUIRED
, `features`.`size`=minor_type: INT
mode: OPTIONAL
}
2021-05-25 15:39:21,066 [1f535658-f840-9f0e-1a7b-21080514bb7b:foreman] TRACE o.a.d.e.s.p.ParquetGroupScanStatistics - check schema path `features`.`size` with major type minor_type: INT
mode: OPTIONAL

 current partitionColTypeMap = {`features`.`indices`.`list`.`element`=null, `features`.`type`=minor_type: TINYINT
mode: REQUIRED
, `features`.`size`=minor_type: INT
mode: OPTIONAL
}

So that there is some condition major type is null, if drill use this code, it will catch NullPointerException error:

TypeProtos.MajorType majorType = columnMetadata != null ? columnMetadata.majorType() : null; # 121
!partitionColTypeMap.get(schemaPath).equals(type) # 189

we need to change null to org.apache.drill.common.types.Types.NULL to avoid NullPointerException error