Adrian Ionescu created SPARK-22665:
--------------------------------------
Summary: Dataset API: .repartition() inconsistency / issue
Key: SPARK-22665
URL: https://issues.apache.org/jira/browse/SPARK-22665
Project: Spark
Issue Type: Improvement
Components: SQL
Affects Versions: 2.2.0
Reporter: Adrian Ionescu
We currently have two functions for explicitly repartitioning a Dataset:
{code}
def repartition(numPartitions: Int)
{code}
and
{code}
def repartition(numPartitions: Int, partitionExprs: Column*)
{code}
The second function's signature allows it to be called with an empty list of
expressions as well.
However:
* {{df.repartition(numPartitions)}} does RoundRobin partitioning
* {{df.repartition(numPartitions, Seq.empty: _*)}} does HashPartitioning on a
constant, effectively moving all tuples to a single partition
Not only is this inconsistent, but the latter behavior is very undesirable: it
may hide problems in small-scale prototype code, but will inevitably fail (or
have terrible performance) in production.
I suggest we should make it:
- either throw an {{IllegalArgumentException}}
- or do RoundRobin partitioning, just like {{df.repartition(numPartitions)}}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
