Matthew Jacobs has posted comments on this change. Change subject: IMPALA-5757: Order-dependent comparison fails in test_kudu ......................................................................
Patch Set 1: > > > > (1 comment) > > > > > > I recently changed several places in the FE to use > LinkedHashMaps > > > in order to get deterministic ordering. That included > > > ToSqlUtils.java:212, so I'm surprised the order is not > > > deterministic. I suspect that file has the code that needs > > fixing, > > > possibly L427 propertyMapToSql(). Maybe it'd be best to sort > the > > > properties to be completely independent from the underlying > > > storage. > > > > Yup, that looks correct. I think we can just create a > LinkedHashSet > > from the entrySet(). Hopefully that won't break too many existing > > tests. > > At that poing propertyMap is already a LinkedHashMap, so I think > the order of the entrySet should be stable (see > https://stackoverflow.com/questions/1190083/does-entryset-in-a-linkedhashmap-also-guarantee-order). > > I think msTable.getParameters() returns a Map<>, which means we may > have to sort it to get a deterministic order. Yeah I was just realizing that as well. Thanks -- To view, visit http://gerrit.cloudera.org:8080/7575 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I64ce0c49fcbab3a4810d52c131bc176f8567ff9e Gerrit-PatchSet: 1 Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-Owner: Matthew Jacobs <[email protected]> Gerrit-Reviewer: Jim Apple <[email protected]> Gerrit-Reviewer: Lars Volker <[email protected]> Gerrit-Reviewer: Matthew Jacobs <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: No
