cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Cody Rank (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7800) Incomplete results when selecting a particular column in the opposite order of my clustering
Date Wed, 20 Aug 2014 23:21:29 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7800?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14104782#comment-14104782
] 

Cody Rank commented on CASSANDRA-7800:
--------------------------------------

When I connect with cqlsh, I get a bunch of version numbers:
[cqlsh 4.1.1 | Cassandra 2.0.8 | CQL spec 3.1.1 | Thrift protocol 19.39.0]

It's a single-node cluster running on Amazon's EC2. I dumped the table to a text file with
the COPY command, and imported it back in as DeviceCounterMonth2 and was able to reproduce
it against the new table.

> Incomplete results when selecting a particular column in the opposite order of my clustering
> --------------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-7800
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7800
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>         Environment: cqlsh 4.1.1 | Cassandra 2.0.8 | CQL spec 3.1.1 | Thrift protocol
19.39.0
>            Reporter: Cody Rank
>         Attachments: dcm.txt
>
>
> When I run the following query, I get 96 rows back:
> {code}
> SELECT * FROM "DeviceCounterMonth" WHERE "counterGuid" = 3ae09592-91b3-568f-93d0-d7a77d46d2d7
AND "deviceGuid" = ae2fc84d-b85b-4fcf-9881-71af4cc5419d AND "startOfMonthLocal" = '2014-08-01T00:00:00+00:00'
ORDER BY "sampleBucketTimeLocal" DESC;
> {code}
> However, If I change DESC to ASC, I only get back a single row. In that row, {{sampleBucketTimeLocal}}
is {{null}}, which is not true of any of the rows returned by the first query, and should
be impossible since it's part of the primary key.
> Further, if I select specific columns (instead of *), as long as I leave out tenantGuid,
the query returns the same expected 96 rows regardless of whether I use DESC or ASC.
> I haven't been able to create a minimal repro, so I'm attaching a dump of the table.
The schema is as follows:
> {code}
> CREATE TABLE "DeviceCounterMonth" (
>   "counterGuid" uuid,
>   "deviceGuid" uuid,
>   "startOfMonthLocal" text,
>   "sampleBucketTimeLocal" text,
>   max float,
>   mean float,
>   min float,
>   samples bigint,
>   "stdDevs" float,
>   "tenantGuid" uuid,
>   PRIMARY KEY (("counterGuid", "deviceGuid", "startOfMonthLocal"), "sampleBucketTimeLocal")
> ) WITH CLUSTERING ORDER BY ("sampleBucketTimeLocal" DESC) AND
>   bloom_filter_fp_chance=0.010000 AND
>   caching='KEYS_ONLY' AND
>   comment='' AND
>   dclocal_read_repair_chance=0.000000 AND
>   gc_grace_seconds=864000 AND
>   index_interval=128 AND
>   read_repair_chance=0.100000 AND
>   populate_io_cache_on_flush='false' AND
>   default_time_to_live=0 AND
>   speculative_retry='99.0PERCENTILE' AND
>   memtable_flush_period_in_ms=0 AND
>   compaction={'class': 'SizeTieredCompactionStrategy'} AND
>   compression={'sstable_compression': 'LZ4Compressor'};
> {code}



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message