[
https://issues.apache.org/jira/browse/SPARK-50839?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17913531#comment-17913531
]
Hyukjin Kwon commented on SPARK-50839:
--------------------------------------
Actually I think we should just leave this behavior as is ... maybe we had to
allow this in Spark Classic too.. Spark Connect is already out, and I want to
avoid breaking changes. Let's leave as is for now.
> Spark Connect createDataFrame behavior does not match with classic Spark
> ------------------------------------------------------------------------
>
> Key: SPARK-50839
> URL: https://issues.apache.org/jira/browse/SPARK-50839
> Project: Spark
> Issue Type: Bug
> Components: Connect, PySpark
> Affects Versions: 4.0.0
> Reporter: Harsh Motwani
> Priority: Major
>
> Classic Spark and Spark Connect behave differently when `List(Integer)` is
> passed as the first argument instead of `List(Row)` or `List(Tuple)`. I
> personally agree with the classic Spark behavior and Spark Connect should
> fail in this case as well.
> In classic Spark shell (PySpark):
> {code:java}
> >>> spark.createDataFrame([1], "v int").show()
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> ...
> pyspark.errors.exceptions.base.PySparkTypeError:
> [FIELD_DATA_TYPE_UNACCEPTABLE] StructType([StructField('v', IntegerType(),
> True)]) can not accept object 1 in type <class 'int'>.
> {code}
> In Spark Connect shell (PySpark):
> {code:java}
> >>> spark.createDataFrame([1], "v int").show()
> 25/01/15 14:05:01 WARN CheckAllocator: More than one DefaultAllocationManager
> on classpath. Choosing first found
> +---+
> | v|
> +---+
> | 1|
> +---+
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
