[jira] [Commented] (SPARK-11437) createDataFrame shouldn't .take() when provided schema
[
https://issues.apache.org/jira/browse/SPARK-11437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15074733#comment-15074733
]
Maciej Bryński commented on SPARK-11437:
There is such an API, but not public.
{code}
from pyspark.sql.types import _verify_type
{code}
> createDataFrame shouldn't .take() when provided schema
> --
>
> Key: SPARK-11437
> URL: https://issues.apache.org/jira/browse/SPARK-11437
> Project: Spark
> Issue Type: Improvement
> Components: PySpark
>Reporter: Jason White
>Assignee: Jason White
> Fix For: 1.6.0
>
>
> When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls
> `.take(10)` to verify the first 10 rows of the RDD match the provided schema.
> Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue
> affected cases where a schema was not provided.
> Verifying the first 10 rows is of limited utility and causes the DAG to be
> executed non-lazily. If necessary, I believe this verification should be done
> lazily on all rows. However, since the caller is providing a schema to
> follow, I think it's acceptable to simply fail if the schema is incorrect.
> https://github.com/apache/spark/blob/master/python/pyspark/sql/context.py#L321-L325
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11437) createDataFrame shouldn't .take() when provided schema
[ https://issues.apache.org/jira/browse/SPARK-11437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15074726#comment-15074726 ] Davies Liu commented on SPARK-11437: [~maver1ck] Maybe we should provide a API to verify the schema (called manually by user)? > createDataFrame shouldn't .take() when provided schema > -- > > Key: SPARK-11437 > URL: https://issues.apache.org/jira/browse/SPARK-11437 > Project: Spark > Issue Type: Improvement > Components: PySpark >Reporter: Jason White >Assignee: Jason White > Fix For: 1.6.0 > > > When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls > `.take(10)` to verify the first 10 rows of the RDD match the provided schema. > Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue > affected cases where a schema was not provided. > Verifying the first 10 rows is of limited utility and causes the DAG to be > executed non-lazily. If necessary, I believe this verification should be done > lazily on all rows. However, since the caller is providing a schema to > follow, I think it's acceptable to simply fail if the schema is incorrect. > https://github.com/apache/spark/blob/master/python/pyspark/sql/context.py#L321-L325 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11437) createDataFrame shouldn't .take() when provided schema
[
https://issues.apache.org/jira/browse/SPARK-11437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15068627#comment-15068627
]
Maciej Bryński commented on SPARK-11437:
[~davies]
Are you sure that this patch is OK ?
Right now if I'm creating DataFrame from RDD of Rows there is no schema
validation.
So we can create schema with wrong types.
{code}
from pyspark.sql.types import *
schema = StructType([StructField("id", IntegerType()), StructField("name",
IntegerType())])
sqlCtx.createDataFrame(sc.parallelize([Row(id=1, name="abc")]),
schema).collect()
[Row(id=1, name=None)]
{code}
Even better. Column can change places.
{code}
from pyspark.sql.types import *
schema = StructType([StructField("name", IntegerType())])
sqlCtx.createDataFrame(sc.parallelize([Row(id=1, name="abc")]),
schema).collect()
[Row(name=1)]
{code}
> createDataFrame shouldn't .take() when provided schema
> --
>
> Key: SPARK-11437
> URL: https://issues.apache.org/jira/browse/SPARK-11437
> Project: Spark
> Issue Type: Improvement
> Components: PySpark
>Reporter: Jason White
>Assignee: Jason White
> Fix For: 1.6.0
>
>
> When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls
> `.take(10)` to verify the first 10 rows of the RDD match the provided schema.
> Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue
> affected cases where a schema was not provided.
> Verifying the first 10 rows is of limited utility and causes the DAG to be
> executed non-lazily. If necessary, I believe this verification should be done
> lazily on all rows. However, since the caller is providing a schema to
> follow, I think it's acceptable to simply fail if the schema is incorrect.
> https://github.com/apache/spark/blob/master/python/pyspark/sql/context.py#L321-L325
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
-
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11437) createDataFrame shouldn't .take() when provided schema
[ https://issues.apache.org/jira/browse/SPARK-11437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14984166#comment-14984166 ] Apache Spark commented on SPARK-11437: -- User 'JasonMWhite' has created a pull request for this issue: https://github.com/apache/spark/pull/9392 > createDataFrame shouldn't .take() when provided schema > -- > > Key: SPARK-11437 > URL: https://issues.apache.org/jira/browse/SPARK-11437 > Project: Spark > Issue Type: Improvement > Components: PySpark >Reporter: Jason White > > When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls > `.take(10)` to verify the first 10 rows of the RDD match the provided schema. > Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue > affected cases where a schema was not provided. > Verifying the first 10 rows is of limited utility and causes the DAG to be > executed non-lazily. If necessary, I believe this verification should be done > lazily on all rows. However, since the caller is providing a schema to > follow, I think it's acceptable to simply fail if the schema is incorrect. > https://github.com/apache/spark/blob/master/python/pyspark/sql/context.py#L321-L325 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-11437) createDataFrame shouldn't .take() when provided schema
[ https://issues.apache.org/jira/browse/SPARK-11437?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14984009#comment-14984009 ] Jason White commented on SPARK-11437: - [~marmbrus] We briefly discussed this at SparkSummitEU this week. > createDataFrame shouldn't .take() when provided schema > -- > > Key: SPARK-11437 > URL: https://issues.apache.org/jira/browse/SPARK-11437 > Project: Spark > Issue Type: Improvement > Components: PySpark >Reporter: Jason White > > When creating a DataFrame from an RDD in PySpark, `createDataFrame` calls > `.take(10)` to verify the first 10 rows of the RDD match the provided schema. > Similar to https://issues.apache.org/jira/browse/SPARK-8070, but that issue > affected cases where a schema was not provided. > Verifying the first 10 rows is of limited utility and causes the DAG to be > executed non-lazily. If necessary, I believe this verification should be done > lazily on all rows. However, since the caller is providing a schema to > follow, I think it's acceptable to simply fail if the schema is incorrect. > https://github.com/apache/spark/blob/master/python/pyspark/sql/context.py#L321-L325 -- This message was sent by Atlassian JIRA (v6.3.4#6332) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
