[ https://issues.apache.org/jira/browse/CASSANDRA-13262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15906894#comment-15906894 ]
Murukesh Mohanan commented on CASSANDRA-13262: ---------------------------------------------- This is on the Python side, specifically because the results are converted to an OrderedDict ([bin/cqlsh.py#L500|https://github.com/apache/cassandra/blob/trunk/bin/cqlsh.py#L500]): {code} self.session.row_factory = ordered_dict_factory {code} Dictionaries of course don't support duplicate keys. The default row_factory is a named tuple, which also doesn't like duplicate keys, so we have changes to the key names: {code} Row(rack=u'rack1', timeout=5000, rack_=u'rack1') OrderedDict([(u'rack', u'rack1'), (u'timeout', 5000)]) {code} The simple fix would be explicitly list the values corresponding to each column in [print_static_result()|https://github.com/apache/cassandra/blob/trunk/bin/cqlsh.py#L1115]: {code} formatted_values = [map(self.myformat_value, [row[c] for c in column_names], cql_types) for row in result.current_rows] {code} And that sort of negates the point of using an OrderedDict in the first place. > Incorrect cqlsh results when selecting same columns multiple times > ------------------------------------------------------------------ > > Key: CASSANDRA-13262 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13262 > Project: Cassandra > Issue Type: Bug > Reporter: Stefan Podkowinski > Priority: Minor > Labels: lhf > > Just stumbled over this on trunk: > {quote} > cqlsh:test1> select a, b, c from table1; > a | b | c > ---+------+----- > 1 | b | 2 > 2 | null | 2.2 > (2 rows) > cqlsh:test1> select a, a, b, c from table1; > a | a | b | c > ---+------+-----+------ > 1 | b | 2 | null > 2 | null | 2.2 | null > (2 rows) > cqlsh:test1> select a, a, a, b, c from table1; > a | a | a | b | c > ---+------+---------------+------+------ > 1 | b | 2.0 | null | null > 2 | null | 2.20000004768 | null | null > {quote} > My guess is that his is on the Python side, but haven't really looked into it. -- This message was sent by Atlassian JIRA (v6.3.15#6346)