[
https://issues.apache.org/jira/browse/SPARK-2890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14116169#comment-14116169
]
Yin Huai commented on SPARK-2890:
---------------------------------
Do we also want to handle ambiguity introduced by join?
For example,
{code:sql}
CREATE TABLE t1 (key int, value string);
CREATE TABLE t2 (key int, value string);
SELECT x.key, x.value, y.key, y.value FROM t1 LEFT OUTER JOIN t2 ON (t1.key =
t2.key);
{code}
For this kind of cases, there will be an exception when users try to get the
schema of the result SchemaRDD. I feel it is better to keep the exception and
ask users to use AS to rename those columns.
> Spark SQL should allow SELECT with duplicated columns
> -----------------------------------------------------
>
> Key: SPARK-2890
> URL: https://issues.apache.org/jira/browse/SPARK-2890
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 1.1.0
> Reporter: Jianshi Huang
> Assignee: Michael Armbrust
>
> Spark reported error java.lang.IllegalArgumentException with messages:
> java.lang.IllegalArgumentException: requirement failed: Found fields with the
> same name.
> at scala.Predef$.require(Predef.scala:233)
> at
> org.apache.spark.sql.catalyst.types.StructType.<init>(dataTypes.scala:317)
> at
> org.apache.spark.sql.catalyst.types.StructType$.fromAttributes(dataTypes.scala:310)
> at
> org.apache.spark.sql.parquet.ParquetTypesConverter$.convertToString(ParquetTypes.scala:306)
> at
> org.apache.spark.sql.parquet.ParquetTableScan.execute(ParquetTableOperations.scala:83)
> at
> org.apache.spark.sql.execution.Filter.execute(basicOperators.scala:57)
> at
> org.apache.spark.sql.execution.SparkPlan.executeCollect(SparkPlan.scala:85)
> at org.apache.spark.sql.SchemaRDD.collect(SchemaRDD.scala:433)
> After trial and error, it seems it's caused by duplicated columns in my
> select clause.
> I made the duplication on purpose for my code to parse correctly. I think we
> should allow users to specify duplicated columns as return value.
> Jianshi
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
