[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14172630#comment-14172630 ] Constance Eustace commented on CASSANDRA-4762: -- Ok, thanks. I can use a different schema, and maybe try out the new one. I'm pretty interested in Presto / Shark adhoc queries, that's the main motivation of the reengineer... this ticket is for supporting our main lookup use cases... Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14171446#comment-14171446 ] Tyler Hobbs commented on CASSANDRA-4762: bq. So does 3.0 drop this back another 6 months? Yes. I cannot make any guarantees about deadlines, but we should at least have a beta or release candidate available in 6 months. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14167562#comment-14167562 ] Constance Eustace commented on CASSANDRA-4762: -- So does 3.0 drop this back another 6 months? I'm planning a large schema reengineer and need to know when I can rely on it being there... Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14137152#comment-14137152 ] Benjamin Lerer commented on CASSANDRA-4762: --- Pushed the fix to [branch|https://github.com/blerer/cassandra/commits/CASSANDRA-4762]. I have fixed all the points but I am not fully happy with the new name for the add method of CompositesBuilder: {quote}addElementToAll(ByteBuffer){quote} and {quote}addEachElementToAll(ListByteBuffer){quote}. I am willing to take any suggestion. I tried to make the javadoc explicite by providing some examples to compensate for those not so clear names. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14135188#comment-14135188 ] Benjamin Lerer commented on CASSANDRA-4762: --- The branch for 3.0 is: [branch|https://github.com/blerer/cassandra/commits/CASSANDRA-4762] Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14136104#comment-14136104 ] Tyler Hobbs commented on CASSANDRA-4762: Review comments * {{SelectStatement.buildBound()}}: ** In the IN case, we no longer remove duplicates (TreeSet did this automatically) ** The single-item and multi-item cases could be combined in the current version (but I guess not if you add duplicate removal) * {{CompositesBuilder}} ** No curly braces needed for single-line if statements ({{add()}}, {{buildWithEOC()}}) ** {{add(ListByteBuffer)}} is a little deceptive. I would have guessed that it was equivalent to calling {{.add(ByteBuffer)}} for each value in the list, but that's not the case. Can you come up with a clearer name? * Add a few more tests: ** Tables that use a mixture of ASC/DESC clustering orders (in combination with ordering clauses) ** Duplicate items in the list of IN values ** Tests like this one, but with all inequality operators ({{}}, {{=}}, {{}}, {{=}}): {{select * from %s where a = ? and c = ? and b in (?, ?)}} Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14129031#comment-14129031 ] Tyler Hobbs commented on CASSANDRA-4762: Since this is a new feature and affects a fairly core part of query handling, I would prefer to keep this targeted for 3.0 instead of 2.1. We want to stabilize 2.1 as quickly as possible. What do you think, [~slebresne]? Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 2.1.1 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14129040#comment-14129040 ] Sylvain Lebresne commented on CASSANDRA-4762: - Yes, it's probably safer to push to 3.0 indeed. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 2.1.1 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14123924#comment-14123924 ] Michael Shuler commented on CASSANDRA-4762: --- Tested out your 2.1 branch - builds fine, the new unit test passes for me, and I enabled and ran the 3 dtests associated with this ticket number: {noformat} (master *)mshuler@hana:~/git/cassandra-dtest$ nosetests -vs --test-select-config=../run_excluded_tests.cfg cql_tests.py multi_in_compact_non_composite_test (cql_tests.TestCQL) ... ok multi_in_compact_test (cql_tests.TestCQL) ... ok multi_in_test (cql_tests.TestCQL) ... ok {noformat} Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 2.1.1 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14119511#comment-14119511 ] Peter Mädel commented on CASSANDRA-4762: any chances to get this fix into 2.1.x ? Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14119525#comment-14119525 ] Benjamin Lerer commented on CASSANDRA-4762: --- I will cancel this patch and make another one for the 2.1.X. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14120006#comment-14120006 ] Benjamin Lerer commented on CASSANDRA-4762: --- Pushed a new [branch|https://github.com/blerer/cassandra/commits/CASSANDRA-2.1-4762] with the same fix for 2.1. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql, docs Fix For: 2.1.1 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-4762) Support IN clause for any clustering column
[ https://issues.apache.org/jira/browse/CASSANDRA-4762?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14118116#comment-14118116 ] Benjamin Lerer commented on CASSANDRA-4762: --- Pushed [branch|https://github.com/blerer/cassandra/commits/CASSANDRA-4762] with the fix. The fix add a new class CompositesBuilder that can be used to build easily mutiple Composites. Internally it delegate the task of building the Composite to the CBuilder provided in the constructor. SelectStatement has been then modified to use this CompositesBuilder in the buildBound and getRequestedColumns methods. Support IN clause for any clustering column --- Key: CASSANDRA-4762 URL: https://issues.apache.org/jira/browse/CASSANDRA-4762 Project: Cassandra Issue Type: Improvement Components: Core Reporter: T Jake Luciani Assignee: Benjamin Lerer Labels: cql Fix For: 3.0 Attachments: 4762-1.txt Given CASSANDRA-3885 It seems it should be possible to store multiple ranges for many predicates even the inner parts of a composite column. They could be expressed as a expanded set of filter queries. example: {code} CREATE TABLE test ( name text, tdate timestamp, tdate2 timestamp, tdate3 timestamp, num double, PRIMARY KEY(name,tdate,tdate2,tdate3) ) WITH COMPACT STORAGE; SELECT * FROM test WHERE name IN ('a','b') and tdate IN ('2010-01-01','2011-01-01') and tdate2 IN ('2010-01-01','2011-01-01') and tdate3 IN ('2010-01-01','2011-01-01') {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)