Partitioned delta table failing to write checkpoints?-CodePudding

I'm using pyspark with spark 3.2.1 and delta table package ("io.delta:delta-core_2.12:2.1.0").

I have partitioned my table by Code, Year and Month like below:

table.write.partitionBy("Code", "Year", "Month").format('delta').save(path)

When I run a merge on this table:

deltaTable.alias("t0").merge(
        df.alias("t1"),
        " t0.Code= t1.CodeAND "
        " t0.Year = t1.Year AND "
        " t0.Month = t1.Month AND "
        " t0.Date = t1.Date"
    ).whenMatchedUpdate(
        set={
            ...
        }
    ).execute()

I got the following warning:

WARN DAGScheduler: Broadcasting large task binary with size 1861.8 KiB

Then I got the following error after some minutes:

py4j.protocol.Py4JJavaError: An error occurred while calling o2070.execute.
: java.lang.NoSuchMethodError: org.apache.spark.sql.catalyst.expressions.ElementAt$.apply$default$3()Lscala/Option;
    at org.apache.spark.sql.delta.CheckpointV2$.$anonfun$extractPartitionValues$1(Checkpoints.scala:727)
    at scala.collection.TraversableLike.$anonfun$map$1(TraversableLike.scala:286)
    at scala.collection.Iterator.foreach(Iterator.scala:943)
    at scala.collection.Iterator.foreach$(Iterator.scala:943)
    at scala.collection.AbstractIterator.foreach(Iterator.scala:1431)
    at scala.collection.IterableLike.foreach(IterableLike.scala:74)
    at scala.collection.IterableLike.foreach$(IterableLike.scala:73)
    at org.apache.spark.sql.types.StructType.foreach(StructType.scala:102)
    at scala.collection.TraversableLike.map(TraversableLike.scala:286)
    at scala.collection.TraversableLike.map$(TraversableLike.scala:279)
    at org.apache.spark.sql.types.StructType.map(StructType.scala:102)
    at org.apache.spark.sql.delta.CheckpointV2$.extractPartitionValues(Checkpoints.scala:724)
    at org.apache.spark.sql.delta.Checkpoints$.buildCheckpoint(Checkpoints.scala:689)
    at org.apache.spark.sql.delta.Checkpoints$.$anonfun$writeCheckpoint$1(Checkpoints.scala:531)
    at org.apache.spark.sql.delta.metering.DeltaLogging.withDmqTag(DeltaLogging.scala:143)
    at org.apache.spark.sql.delta.metering.DeltaLogging.withDmqTag$(DeltaLogging.scala:142)
    at org.apache.spark.sql.delta.Checkpoints$.withDmqTag(Checkpoints.scala:460)
    at org.apache.spark.sql.delta.Checkpoints$.writeCheckpoint(Checkpoints.scala:487)
    at org.apache.spark.sql.delta.Checkpoints.writeCheckpointFiles(Checkpoints.scala:361)
    at org.apache.spark.sql.delta.Checkpoints.writeCheckpointFiles$(Checkpoints.scala:359)
    at org.apache.spark.sql.delta.DeltaLog.writeCheckpointFiles(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.Checkpoints.checkpointAndCleanUpDeltaLog(Checkpoints.scala:346)
    at org.apache.spark.sql.delta.Checkpoints.checkpointAndCleanUpDeltaLog$(Checkpoints.scala:344)
    at org.apache.spark.sql.delta.DeltaLog.checkpointAndCleanUpDeltaLog(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.Checkpoints.$anonfun$checkpoint$2(Checkpoints.scala:318)
    at scala.runtime.java8.JFunction0$mcV$sp.apply(JFunction0$mcV$sp.java:23)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile(DeltaLogging.scala:139)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile$(DeltaLogging.scala:137)
    at org.apache.spark.sql.delta.DeltaLog.recordFrameProfile(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.metering.DeltaLogging.$anonfun$recordDeltaOperationInternal$1(DeltaLogging.scala:132)
    at com.databricks.spark.util.DatabricksLogging.recordOperation(DatabricksLogging.scala:77)
    at com.databricks.spark.util.DatabricksLogging.recordOperation$(DatabricksLogging.scala:67)
    at org.apache.spark.sql.delta.DeltaLog.recordOperation(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperationInternal(DeltaLogging.scala:131)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation(DeltaLogging.scala:121)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation$(DeltaLogging.scala:109)
    at org.apache.spark.sql.delta.DeltaLog.recordDeltaOperation(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.Checkpoints.$anonfun$checkpoint$1(Checkpoints.scala:314)
    at scala.runtime.java8.JFunction0$mcV$sp.apply(JFunction0$mcV$sp.java:23)
    at org.apache.spark.sql.delta.metering.DeltaLogging.withDmqTag(DeltaLogging.scala:143)
    at org.apache.spark.sql.delta.metering.DeltaLogging.withDmqTag$(DeltaLogging.scala:142)
    at org.apache.spark.sql.delta.DeltaLog.withDmqTag(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.Checkpoints.checkpoint(Checkpoints.scala:313)
    at org.apache.spark.sql.delta.Checkpoints.checkpoint$(Checkpoints.scala:312)
    at org.apache.spark.sql.delta.DeltaLog.checkpoint(DeltaLog.scala:63)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.postCommit(OptimisticTransaction.scala:1097)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.postCommit$(OptimisticTransaction.scala:1092)
    at org.apache.spark.sql.delta.OptimisticTransaction.postCommit(OptimisticTransaction.scala:101)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.liftedTree1$1(OptimisticTransaction.scala:750)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.$anonfun$commit$1(OptimisticTransaction.scala:691)
    at scala.runtime.java8.JFunction0$mcJ$sp.apply(JFunction0$mcJ$sp.java:23)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile(DeltaLogging.scala:139)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile$(DeltaLogging.scala:137)
    at org.apache.spark.sql.delta.OptimisticTransaction.recordFrameProfile(OptimisticTransaction.scala:101)
    at org.apache.spark.sql.delta.metering.DeltaLogging.$anonfun$recordDeltaOperationInternal$1(DeltaLogging.scala:132)
    at com.databricks.spark.util.DatabricksLogging.recordOperation(DatabricksLogging.scala:77)
    at com.databricks.spark.util.DatabricksLogging.recordOperation$(DatabricksLogging.scala:67)
    at org.apache.spark.sql.delta.OptimisticTransaction.recordOperation(OptimisticTransaction.scala:101)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperationInternal(DeltaLogging.scala:131)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation(DeltaLogging.scala:121)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation$(DeltaLogging.scala:109)
    at org.apache.spark.sql.delta.OptimisticTransaction.recordDeltaOperation(OptimisticTransaction.scala:101)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.commit(OptimisticTransaction.scala:688)
    at org.apache.spark.sql.delta.OptimisticTransactionImpl.commit$(OptimisticTransaction.scala:686)
    at org.apache.spark.sql.delta.OptimisticTransaction.commit(OptimisticTransaction.scala:101)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.$anonfun$run$2(MergeIntoCommand.scala:363)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.$anonfun$run$2$adapted(MergeIntoCommand.scala:319)
    at org.apache.spark.sql.delta.DeltaLog.withNewTransaction(DeltaLog.scala:221)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.$anonfun$run$1(MergeIntoCommand.scala:319)
    at scala.runtime.java8.JFunction0$mcV$sp.apply(JFunction0$mcV$sp.java:23)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile(DeltaLogging.scala:139)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordFrameProfile$(DeltaLogging.scala:137)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.recordFrameProfile(MergeIntoCommand.scala:215)
    at org.apache.spark.sql.delta.metering.DeltaLogging.$anonfun$recordDeltaOperationInternal$1(DeltaLogging.scala:132)
    at com.databricks.spark.util.DatabricksLogging.recordOperation(DatabricksLogging.scala:77)
    at com.databricks.spark.util.DatabricksLogging.recordOperation$(DatabricksLogging.scala:67)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.recordOperation(MergeIntoCommand.scala:215)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperationInternal(DeltaLogging.scala:131)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation(DeltaLogging.scala:121)
    at org.apache.spark.sql.delta.metering.DeltaLogging.recordDeltaOperation$(DeltaLogging.scala:109)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.recordDeltaOperation(MergeIntoCommand.scala:215)
    at org.apache.spark.sql.delta.commands.MergeIntoCommand.run(MergeIntoCommand.scala:317)
    at io.delta.tables.DeltaMergeBuilder.$anonfun$execute$1(DeltaMergeBuilder.scala:230)
    at org.apache.spark.sql.delta.util.AnalysisHelper.improveUnsupportedOpError(AnalysisHelper.scala:104)
    at org.apache.spark.sql.delta.util.AnalysisHelper.improveUnsupportedOpError$(AnalysisHelper.scala:90)
    at io.delta.tables.DeltaMergeBuilder.improveUnsupportedOpError(DeltaMergeBuilder.scala:122)
    at io.delta.tables.DeltaMergeBuilder.execute(DeltaMergeBuilder.scala:206)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
    at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
    at py4j.Gateway.invoke(Gateway.java:282)
    at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
    at py4j.commands.CallCommand.execute(CallCommand.java:79)
    at py4j.ClientServerConnection.waitForCommands(ClientServerConnection.java:182)
    at py4j.ClientServerConnection.run(ClientServerConnection.java:106)
    at java.lang.Thread.run(Thread.java:748)

This error only happens when the table is partitioned! If I not use partitions on my folder my code runs normally.

I'm using this code inside a for loop, performing multiple updates on that table!

So my doubt is: Why is this happening when the table is partitioned?

CodePudding user response：

You're using incompatible version of the Delta Lake - as you can see in the docs, the version 2.1.0 is compatible with Spark 3.3.0, so you can't use it with 3.2.1. You need to take 2.0.1 instead