Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of jbellis@gmail.com designates
 209.85.215.172 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAP5HrRYrnnNsadiPp8981dWT-XZoo6pvqPigwW=y2QThpfd=bA@mail.gmail.com>
References: 
 <CAP5HrRZO8RsAVZTU0iGG19d35xoHKJnfgFQP_RvbD2qZLT59RQ@mail.gmail.com>
 <CALamADLDLoh2aPOCZLKpKM4UyGFx+OgriAv_8VhShrtg9FRk=Q@mail.gmail.com>
 <CAP5HrRYrnnNsadiPp8981dWT-XZoo6pvqPigwW=y2QThpfd=bA@mail.gmail.com>
From: Jonathan Ellis <jbellis@gmail.com>
Date: Tue, 11 Oct 2011 19:54:25 -0500
Message-ID: 
 <CALdd-zjBE=ZaAdU8YRnD7r8k+ybGa9EwmTXeD1iSPCsmBJ_EsQ@mail.gmail.com>
Subject: Re: Operator on secondary indexes in 0.8.x (GTE/LTE)
To: user@cassandra.apache.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

simple, elegant, and less performant than just doing a range scan
without the index. :)

On Tue, Oct 11, 2011 at 4:06 PM, Sasha Dolgy <sdolgy@gmail.com> wrote:
> ah, hadn't even thought of that. =A0simple. =A0elegant.
> cheers.
>
> On Tue, Oct 11, 2011 at 11:01 PM, Jake Luciani <jakers@gmail.com> wrote:
>>
>> This hasn't changed in AFAIK, =A0In Brisk we had the same problem in CFS=
 so
>> we created a=A0sentinel=A0value that all rows shared then it works.
>> CASSANDRA-2915 should fix it.
>> On Tue, Oct 11, 2011 at 4:48 PM, Sasha Dolgy <sdolgy@gmail.com> wrote:
>>>
>>> I was trying to get a range of rows based on a secondary_index that was
>>> defined. =A0Any rows where age was greater than or equal to ... it didn=
't
>>> work. =A0Is this a continued limitation? =A0Did a quick look in JIRA, c=
ouldn't
>>> find anything.
>>> The output from "help get;" on the cli contains the following, which le=
d
>>> me to believe it was a limitation on Cassandra 0.7.x and not on 0.8.x .=
..
>>> get <cf> where <col> <operator> <value> [
>>> =A0 =A0 and <col> <operator> <value> and ...] [limit <limit>];
>>> get <cf> where <col> <operator> <function>(<value>) [
>>> =A0 =A0 and <col> <operator> <function> and ...] [limit <limit>];
>>> - operator: Operator to test the column value with. Supported operators
>>> are
>>> =A0 =3D, >, >=3D, <, <=3D .
>>> =A0 In Cassandra 0.7 at least one =3D operator must be present.
>>> [default@sdo] =A0get user where age >=3D 18;
>>> No indexed columns present in index clause with operator EQ
>>> [default@sdo]=A0 get user where gender =3D 1 and age >=3D 18
>>> (returns results)
>>> Tested this behavior on 0.8.2, 0.8.6 and now 0.8.7 ...
>>> create column family user
>>> =A0 with column_type =3D 'Standard'
>>> =A0 and comparator =3D 'UTF8Type'
>>> =A0 and default_validation_class =3D 'BytesType'
>>> =A0 and key_validation_class =3D 'BytesType'
>>> =A0 and memtable_operations =3D 0.24843749999999998
>>> =A0 and memtable_throughput =3D 53
>>> =A0 and memtable_flush_after =3D 1440
>>> =A0 and rows_cached =3D 0.0
>>> =A0 and row_cache_save_period =3D 0
>>> =A0 and keys_cached =3D 200000.0
>>> =A0 and key_cache_save_period =3D 14400
>>> =A0 and read_repair_chance =3D 1.0
>>> =A0 and gc_grace =3D 864000
>>> =A0 and min_compaction_threshold =3D 4
>>> =A0 and max_compaction_threshold =3D 32
>>> =A0 and replicate_on_write =3D true
>>> =A0 and row_cache_provider =3D 'ConcurrentLinkedHashCacheProvider'
>>> =A0 and column_metadata =3D [
>>> =A0 =A0 {column_name : 'gender',
>>> =A0 =A0 validation_class : LongType,
>>> =A0 =A0 index_name : 'user_gender_idx',
>>> =A0 =A0 index_type : 0},
>>> =A0 =A0 {column_name : 'year',
>>> =A0 =A0 validation_class : LongType,
>>> =A0 =A0 index_name : 'user_year_idx',
>>> =A0 =A0 index_type : 0}];
>>>
>>>
>>> --
>>> Sasha Dolgy
>>> sasha.dolgy@gmail.com
>>
>>
>>
>> --
>> http://twitter.com/tjake
>
>
>
> --
> Sasha Dolgy
> sasha.dolgy@gmail.com
>


--=20
Jonathan Ellis
Project Chair, Apache Cassandra
co-founder of DataStax, the source for professional Cassandra support
http://www.datastax.com