[ 
https://issues.apache.org/jira/browse/FLINK-15975?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

testfixer0 updated FLINK-15975:
-------------------------------
    Description: The test `testMap` in `HiveGenericUDFTest` fails. As you 
suggested, we'd be better to change assertions in test code to cater to the 
indeterminacy.  So, I propose a new proposal.   (was: The test `testMap` in 
`HiveGenericUDFTest` invokes two methods `HiveGenericUDFTest.init` and 
`HiveScalarFunction.eval`.  These two methods depend on `getConversion` and 
`toFlinkObject` in class `HiveInspectors`. When the `inspector` is the instance 
of `MapObjectInspector`, it can return a `HashMap`. However, `HashMap` does not 
guarantee any specific order of entries. Thus, the test can fail due to a 
different iteration order.

In this PR, we propose to use a `LinkedHashMap` instead to guarantee the 
iteration order.)

> Use LinkedHashMap for deterministic iterations
> ----------------------------------------------
>
>                 Key: FLINK-15975
>                 URL: https://issues.apache.org/jira/browse/FLINK-15975
>             Project: Flink
>          Issue Type: Improvement
>          Components: Connectors / Hive
>            Reporter: testfixer0
>            Priority: Minor
>              Labels: pull-request-available
>         Attachments: FLINK-15975-000.patch, FLINK-15975-001.patch
>
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> The test `testMap` in `HiveGenericUDFTest` fails. As you suggested, we'd be 
> better to change assertions in test code to cater to the indeterminacy.  So, 
> I propose a new proposal. 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to