[jira] [Comment Edited] (SPARK-5236) java.lang.ClassCastException: org.apache.spark.sql.catalyst.expressions.MutableAny cannot be cast to org.apache.spark.sql.catalyst.expressions.MutableInt
[
https://issues.apache.org/jira/browse/SPARK-5236?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15936222#comment-15936222
]
Jorge Machado edited comment on SPARK-5236 at 3/22/17 12:52 PM:
[~marmbrus] Hi Michael, so I'm experience the same issue. I'm building a
datasource for Hbase with some custom schema. I'm on 1.6.3
I traced down to GeneratePredicates.scala (r: InternalRow) => p.eval(r) and on
my example it tries to do instanceOf[MutableLong] from a String which fails. I
have a filter on the dataframe and a groupby count
{noformat}
/**
*
* @param schema this is how the row has to look like. The returned value from
the next must match this schema
* @param hBaseRelation
* @param repositoryHistory
* @param timeZoneId
* @param tablePartitionInfo
* @param from
* @param to
*/
class TagValueSparkIterator(val hBaseRelation: HBaseRelation,
val schema: StructType,
val repositoryHistory:
DeviceHistoryRepository,
val timeZoneId: String,
val tablePartitionInfo:
TablePartitionInfo,
val from: Long,
val to: Long) extends
Iterator[InternalRow] {
private val internalItr: ClosableIterator[TagValue[Double]]=
repositoryHistory.scanTagValues(from, to, tablePartitionInfo)
override def hasNext: Boolean = internalItr.hasNext
override def next(): InternalRow = {
val tagValue = internalItr.next()
val instant =
ZonedDateTime.ofInstant(Instant.ofEpochSecond(tagValue.getTimestamp),
ZoneId.of(timeZoneId)).toInstant
val timestamp = Timestamp.from(instant)
InternalRow.fromSeq(Array(tagValue.getTimestamp,tagValue.getGuid,tagValue.getGuid,tagValue.getValue))
val mutableRow = new SpecificMutableRow(schema.fields.map(f=>
f.dataType))
for (i <- schema.fields.indices){
updateMutableRow(i,tagValue,mutableRow, schema(i) )
}
mutableRow
}
def updateMutableRow(i: Int, tagValue: TagValue[Double], row:
SpecificMutableRow, field:StructField): Unit = {
//#TODO this is ugly.
field.name match {
case "Date" =>
row.setLong(i,tagValue.getTimestamp.toLong)
case "Device" =>
row.update(i,UTF8String.fromString(tagValue.getGuid))
case "Tag" =>
row.update(i,UTF8String.fromString(tagValue.getTagName))
case "TagValue" => row.setDouble(i,tagValue.getValue)
}
}
override def toString():String ={
"Iterator for Region Name "+tablePartitionInfo.getRegionName+"
Range:"+from+ "until" + "to"
}
}
{noformat}
Then I get :
{noformat}
Driver stacktrace:
org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in
stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage 0.0 (TID
0, localhost): java.lang.ClassCastException:
org.apache.spark.sql.catalyst.expressions.MutableAny cannot be cast to
org.apache.spark.sql.catalyst.expressions.MutableLong
at
org.apache.spark.sql.catalyst.expressions.SpecificMutableRow.getLong(SpecificMutableRow.scala:301)
at
org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificPredicate.eval(Unknown
Source)
at
org.apache.spark.sql.catalyst.expressions.codegen.GeneratePredicate$$anonfun$create$2.apply(GeneratePredicate.scala:68)
at
org.apache.spark.sql.catalyst.expressions.codegen.GeneratePredicate$$anonfun$create$2.apply(GeneratePredicate.scala:68)
at
org.apache.spark.sql.execution.Filter$$anonfun$2$$anonfun$apply$2.apply(basicOperators.scala:74)
at
org.apache.spark.sql.execution.Filter$$anonfun$2$$anonfun$apply$2.apply(basicOperators.scala:72)
at scala.collection.Iterator$$anon$14.hasNext(Iterator.scala:390)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
{noformat}
was (Author: jomach):
[~marmbrus] Hi Michael, so I'm experience the same issue. I'm building a
datasource for Hbase with some custom schema. I'm on 1.6.3
I traced down to GeneratePredicates.scala (r: InternalRow) => p.eval(r)
{noformat}
/**
*
* @param schema this is how the row has to look like. The returned value from
the next must match this schema
* @param hBaseRelation
* @param repositoryHistory
* @param timeZoneId
* @param tablePartitionInfo
* @param from
* @param to
*/
class TagValueSparkIterator(val hBaseRelation: HBaseRelation,
[jira] [Comment Edited] (SPARK-5236) java.lang.ClassCastException: org.apache.spark.sql.catalyst.expressions.MutableAny cannot be cast to org.apache.spark.sql.catalyst.expressions.MutableInt
[
https://issues.apache.org/jira/browse/SPARK-5236?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15936222#comment-15936222
]
Jorge Machado edited comment on SPARK-5236 at 3/22/17 12:47 PM:
[~marmbrus] Hi Michael, so I'm experience the same issue. I'm building a
datasource for Hbase with some custom schema. I'm on 1.6.3
I traced down to GeneratePredicates.scala (r: InternalRow) => p.eval(r)
{noformat}
/**
*
* @param schema this is how the row has to look like. The returned value from
the next must match this schema
* @param hBaseRelation
* @param repositoryHistory
* @param timeZoneId
* @param tablePartitionInfo
* @param from
* @param to
*/
class TagValueSparkIterator(val hBaseRelation: HBaseRelation,
val schema: StructType,
val repositoryHistory:
DeviceHistoryRepository,
val timeZoneId: String,
val tablePartitionInfo:
TablePartitionInfo,
val from: Long,
val to: Long) extends
Iterator[InternalRow] {
private val internalItr: ClosableIterator[TagValue[Double]]=
repositoryHistory.scanTagValues(from, to, tablePartitionInfo)
override def hasNext: Boolean = internalItr.hasNext
override def next(): InternalRow = {
val tagValue = internalItr.next()
val instant =
ZonedDateTime.ofInstant(Instant.ofEpochSecond(tagValue.getTimestamp),
ZoneId.of(timeZoneId)).toInstant
val timestamp = Timestamp.from(instant)
InternalRow.fromSeq(Array(tagValue.getTimestamp,tagValue.getGuid,tagValue.getGuid,tagValue.getValue))
val mutableRow = new SpecificMutableRow(schema.fields.map(f=>
f.dataType))
for (i <- schema.fields.indices){
updateMutableRow(i,tagValue,mutableRow, schema(i) )
}
mutableRow
}
def updateMutableRow(i: Int, tagValue: TagValue[Double], row:
SpecificMutableRow, field:StructField): Unit = {
//#TODO this is ugly.
field.name match {
case "Date" =>
row.setLong(i,tagValue.getTimestamp.toLong)
case "Device" =>
row.update(i,UTF8String.fromString(tagValue.getGuid))
case "Tag" =>
row.update(i,UTF8String.fromString(tagValue.getTagName))
case "TagValue" => row.setDouble(i,tagValue.getValue)
}
}
override def toString():String ={
"Iterator for Region Name "+tablePartitionInfo.getRegionName+"
Range:"+from+ "until" + "to"
}
}
{noformat}
Then I get :
{noformat}
Driver stacktrace:
org.apache.spark.SparkException: Job aborted due to stage failure: Task 0 in
stage 0.0 failed 1 times, most recent failure: Lost task 0.0 in stage 0.0 (TID
0, localhost): java.lang.ClassCastException:
org.apache.spark.sql.catalyst.expressions.MutableAny cannot be cast to
org.apache.spark.sql.catalyst.expressions.MutableLong
at
org.apache.spark.sql.catalyst.expressions.SpecificMutableRow.getLong(SpecificMutableRow.scala:301)
at
org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificPredicate.eval(Unknown
Source)
at
org.apache.spark.sql.catalyst.expressions.codegen.GeneratePredicate$$anonfun$create$2.apply(GeneratePredicate.scala:68)
at
org.apache.spark.sql.catalyst.expressions.codegen.GeneratePredicate$$anonfun$create$2.apply(GeneratePredicate.scala:68)
at
org.apache.spark.sql.execution.Filter$$anonfun$2$$anonfun$apply$2.apply(basicOperators.scala:74)
at
org.apache.spark.sql.execution.Filter$$anonfun$2$$anonfun$apply$2.apply(basicOperators.scala:72)
at scala.collection.Iterator$$anon$14.hasNext(Iterator.scala:390)
at scala.collection.Iterator$$anon$11.hasNext(Iterator.scala:327)
{noformat}
was (Author: jomach):
[~marmbrus] Hi Michael, so I'm experience the same issue. I'm building a
datasource for Hbase with some custom schema.
{noformat}
/**
*
* @param schema this is how the row has to look like. The returned value from
the next must match this schema
* @param hBaseRelation
* @param repositoryHistory
* @param timeZoneId
* @param tablePartitionInfo
* @param from
* @param to
*/
class TagValueSparkIterator(val hBaseRelation: HBaseRelation,
val schema: StructType,
val repositoryHistory:
DeviceHistoryRepository,
val timeZoneId: String,
