[
https://issues.apache.org/jira/browse/SPARK-36984?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lukáš updated SPARK-36984:
--------------------------
Description:
The documentation at
[https://spark.apache.org/docs/latest/streaming-programming-guide.html#advanced-sources]
clearly states that *Kafka* (and Kinesis) are available in the Python API v
3.1.2 in *Spark Streaming (DStreams)*. (see attachments for highlight)
However, there is no way to create DStream from Kafka in PySpark >= 3.0.0, as
the `kafka.py` file is missing in
[https://github.com/apache/spark/tree/master/python/pyspark/streaming]. I'm
coming from PySpark 2.4.4 where this was possible. _Should Kafka be excluded as
advanced source for spark streaming in Python API in the docs?_
Note that I'm aware of this Kafka integration guide
[https://spark.apache.org/docs/latest/structured-streaming-kafka-integration.html]
but I'm not interested in Structured Streaming as it doesn't support arbitrary
stateful operations in Python. DStreams support this functionality with
`updateStateByKey`.
was:
The documentation at
[https://spark.apache.org/docs/latest/streaming-programming-guide.html#advanced-sources]
clearly states that *Kafka* (and Kinesis) are available in the Python API v
3.1.2 in *Spark Streaming (DStreams)*.
However, there is no way to create DStream from Kafka in PySpark >= 3.0.0, as
the `kafka.py` file is missing in
[https://github.com/apache/spark/tree/master/python/pyspark/streaming]. I'm
coming from PySpark 2.4.4 where this was possible. _Should Kafka be excluded as
advanced source for spark streaming in Python API in the docs?_
Note that I'm aware of this Kafka integration guide
[https://spark.apache.org/docs/latest/structured-streaming-kafka-integration.html]
but I'm not interested in Structured Streaming as it doesn't support arbitrary
stateful operations in Python. DStreams support this functionality with
`updateStateByKey`.
!image-2021-10-12-10-04-03-232.png!
> Misleading Spark Streaming source documentation
> -----------------------------------------------
>
> Key: SPARK-36984
> URL: https://issues.apache.org/jira/browse/SPARK-36984
> Project: Spark
> Issue Type: Documentation
> Components: Documentation, PySpark, Structured Streaming
> Affects Versions: 3.0.1, 3.0.2, 3.0.3, 3.1.0, 3.1.1, 3.1.2
> Reporter: Lukáš
> Priority: Trivial
> Attachments: docs_highlight.png
>
>
> The documentation at
> [https://spark.apache.org/docs/latest/streaming-programming-guide.html#advanced-sources]
> clearly states that *Kafka* (and Kinesis) are available in the Python API v
> 3.1.2 in *Spark Streaming (DStreams)*. (see attachments for highlight)
> However, there is no way to create DStream from Kafka in PySpark >= 3.0.0, as
> the `kafka.py` file is missing in
> [https://github.com/apache/spark/tree/master/python/pyspark/streaming]. I'm
> coming from PySpark 2.4.4 where this was possible. _Should Kafka be excluded
> as advanced source for spark streaming in Python API in the docs?_
>
> Note that I'm aware of this Kafka integration guide
> [https://spark.apache.org/docs/latest/structured-streaming-kafka-integration.html]
> but I'm not interested in Structured Streaming as it doesn't support
> arbitrary stateful operations in Python. DStreams support this functionality
> with `updateStateByKey`.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
