[SPARK-17477] SparkSQL cannot handle schema evolution from Int -> Long when parquet files have Int as its type while hive metastore has Long as its type

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 2.0.0
• Component/s: SQL
• Assignee: Unassigned
• Reporter: Gang Wu
• Created: 09/Sep/16 18:49
• Updated: 12/Dec/22 18:11
• Resolved: 04/Feb/17 17:20

Description

When using SparkSession to read a Hive table which is stored as parquet files. If there has been a schema evolution from int to long of a column. There are some old parquet files use int for the column while some new parquet files use long. In Hive metastore, the type is long (bigint).

Therefore when I use the following:

sparkSession.sql("select * from table").show()

I got the following exception:

16/08/29 17:50:20 WARN scheduler.TaskSetManager: Lost task 0.0 in stage 3.0 (TID 91, XXX): org.apache.parquet.io.ParquetDecodingException: Can not read value at 0 in block 0 in file hdfs://path/to/parquet/1-part-r-00000-d8e4f5aa-b6b9-4cad-8432-a7ae7a590a93.gz.parquet
at org.apache.parquet.hadoop.InternalParquetRecordReader.nextKeyValue(InternalParquetRecordReader.java:228)
at org.apache.parquet.hadoop.ParquetRecordReader.nextKeyValue(ParquetRecordReader.java:201)
at org.apache.spark.sql.execution.datasources.RecordReaderIterator.hasNext(RecordReaderIterator.scala:36)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:408)
at org.apache.spark.sql.execution.datasources.FileScanRDD$$anon$1.hasNext(FileScanRDD.scala:91)
at org.apache.spark.sql.execution.datasources.FileScanRDD$$anon$1.nextIterator(FileScanRDD.scala:128)
at org.apache.spark.sql.execution.datasources.FileScanRDD$$anon$1.hasNext(FileScanRDD.scala:91)
at org.apache.spark.sql.catalyst.expressions.GeneratedClass$GeneratedIterator.processNext(Unknown Source)
at org.apache.spark.sql.execution.BufferedRowIterator.hasNext(BufferedRowIterator.java:43)
at org.apache.spark.sql.execution.WholeStageCodegenExec$$anonfun$8$$anon$1.hasNext(WholeStageCodegenExec.scala:370)
at org.apache.spark.sql.execution.SparkPlan$$anonfun$4.apply(SparkPlan.scala:246)
at org.apache.spark.sql.execution.SparkPlan$$anonfun$4.apply(SparkPlan.scala:240)
at org.apache.spark.rdd.RDD$$anonfun$mapPartitionsInternal$1$$anonfun$apply$24.apply(RDD.scala:784)
at org.apache.spark.rdd.RDD$$anonfun$mapPartitionsInternal$1$$anonfun$apply$24.apply(RDD.scala:784)
at org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
at org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:319)
at org.apache.spark.rdd.RDD.iterator(RDD.scala:283)
at org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:70)
at org.apache.spark.scheduler.Task.run(Task.scala:85)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:274)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)
Caused by: java.lang.ClassCastException: org.apache.spark.sql.catalyst.expressions.MutableLong cannot be cast to org.apache.spark.sql.catalyst.expressions.MutableInt
at org.apache.spark.sql.catalyst.expressions.SpecificMutableRow.setInt(SpecificMutableRow.scala:246)
at org.apache.spark.sql.execution.datasources.parquet.ParquetRowConverter$RowUpdater.setInt(ParquetRowConverter.scala:161)
at org.apache.spark.sql.execution.datasources.parquet.ParquetPrimitiveConverter.addInt(ParquetRowConverter.scala:85)
at org.apache.parquet.column.impl.ColumnReaderImpl$2$3.writeValue(ColumnReaderImpl.java:249)
at org.apache.parquet.column.impl.ColumnReaderImpl.writeCurrentValueToConverter(ColumnReaderImpl.java:365)
at org.apache.parquet.io.RecordReaderImplementation.read(RecordReaderImplementation.java:405)
at org.apache.parquet.hadoop.InternalParquetRecordReader.nextKeyValue(InternalParquetRecordReader.java:209)
... 22 more

But this kind of schema evolution (int => long) is valid is Hive and Presto.

Issue Links

• duplicates: SPARK-16544 Support for conversion from compatible schema for Parquet data source when data types are not matched
• links to: [Github] Pull Request #15264 (wgtmac)