[ 
https://issues.apache.org/jira/browse/FLINK-3665?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15245793#comment-15245793
 ] 

ASF GitHub Bot commented on FLINK-3665:
---------------------------------------

Github user fhueske commented on a diff in the pull request:

    https://github.com/apache/flink/pull/1848#discussion_r60071991
  
    --- Diff: 
flink-java/src/main/java/org/apache/flink/api/java/operators/PartitionOperator.java
 ---
    @@ -98,6 +102,23 @@ public PartitionOperator(DataSet<T> input, Keys<T> 
pKeys, Partitioner<?> customP
                this.customPartitioner = customPartitioner;
                this.distribution = distribution;
        }
    +
    +   /**
    +    * Sets the order of keys for range partitioning.
    +    * NOTE: Only valid for {@link PartitionMethod.RANGE}.
    +    *
    +    * @param orders array of orders for each specified partition key
    +    * @return The partitioneOperator with properly set orders for given 
keys
    +     */
    +   @PublicEvolving
    +   public PartitionOperator<T> withOrders(Order... orders) {
    +           Preconditions.checkState(pMethod == PartitionMethod.RANGE, 
"Orders cannot be applied for %s partition " +
    +                           "method", pMethod);
    +           
Preconditions.checkArgument(pKeys.getOriginalKeyFieldTypes().length == 
orders.length);
    --- End diff --
    
    Can you add a descriptive error message here?


> Range partitioning lacks support to define sort orders
> ------------------------------------------------------
>
>                 Key: FLINK-3665
>                 URL: https://issues.apache.org/jira/browse/FLINK-3665
>             Project: Flink
>          Issue Type: Improvement
>          Components: DataSet API
>    Affects Versions: 1.0.0
>            Reporter: Fabian Hueske
>             Fix For: 1.1.0
>
>
> {{DataSet.partitionByRange()}} does not allow to specify the sort order of 
> fields. This is fine if range partitioning is used to reduce skewed 
> partitioning. 
> However, it is not sufficient if range partitioning is used to sort a data 
> set in parallel. 
> Since {{DataSet.partitionByRange()}} is {{@Public}} API and cannot be easily 
> changed, I propose to add a method {{withOrders(Order... orders)}} to 
> {{PartitionOperator}}. The method should throw an exception if the 
> partitioning method of {{PartitionOperator}} is not range partitioning.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to