[
https://issues.apache.org/jira/browse/CASSANDRA-838?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jonathan Ellis resolved CASSANDRA-838.
--------------------------------------
Resolution: Invalid
If thrift offered such a structure, both would use a OrderedMap<string,
list<COSC>>.
But it doesn't, so we use Map where ordering doesn't matter to give convenient
random access, and List where ordering does matter, so the user has to turn it
into a Map (or OrderedMap) manually.
> Return type of get_slice and multiget_slice is inconsistent with return type
> of get_range_slices
> ------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-838
> URL: https://issues.apache.org/jira/browse/CASSANDRA-838
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Affects Versions: 0.6
> Reporter: Nicholas Telford
> Priority: Minor
>
> As of 0.6.0-beta2, the return type for get_slice and multiget_slice are
> defined as: list<ColumnOrSuperColumn> and map<string,
> list<ColumnOrSuperColumn>> respectively.
> The return type of get_range_slices, is defined as: list<KeySlice>
> KeySlice encapsulates a slice (the "key" for the slice and it's
> ColumnOrSuperColumns).
> This is especially odd since multiget_slice and get_range_slices are
> essentially the same operation, with one getting slices for a specific list
> of keys and the other getting slices for a range of keys. While both return
> types are acceptable ways of representing slices, it would be more consistent
> to just use one - I'm going to assume KeySlice is preferable.
> I'd like to suggest that get_slice and multiget_slice be modified so that
> their return types are: KeySlice and list<KeySlice> respectively.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
