new thrift API

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

new thrift API

Jonas Bonér-2-3
Hey guys.

Is the new API stabilizing?

How is the new range functions suppose to work?
    public List<Column> get_slice_by_names(String keyspace, String
key, ColumnParent column_parent, List<byte[]> column_names, int
consistency_level) throws InvalidRequestException, NotFoundException,
TException;

    public List<Column> get_slice(String keyspace, String key,
ColumnParent column_parent, byte[] start, byte[] finish, boolean
is_ascending, int count, int consistency_level) throws
InvalidRequestException, NotFoundException, TException;

These both have a start, finish and count. I was expecting start and
offset or only start, finish.

Is the new API documented yet?

Thanks.

--
Jonas Bonér

twitter: @jboner
blog:    http://jonasboner.com
work:   http://crisp.se
work:   http://scalablesolutions.se
code:   http://github.com/jboner
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonathan Ellis-3
Stabilizing but not quite finished (329 and 311 are still waiting for
review).  But the fundamentals are the same.

You have start/finish because that's what defines a range.  You have
count because you often want the First N results.

-Jonathan

On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:

> Hey guys.
>
> Is the new API stabilizing?
>
> How is the new range functions suppose to work?
>    public List<Column> get_slice_by_names(String keyspace, String
> key, ColumnParent column_parent, List<byte[]> column_names, int
> consistency_level) throws InvalidRequestException, NotFoundException,
> TException;
>
>    public List<Column> get_slice(String keyspace, String key,
> ColumnParent column_parent, byte[] start, byte[] finish, boolean
> is_ascending, int count, int consistency_level) throws
> InvalidRequestException, NotFoundException, TException;
>
> These both have a start, finish and count. I was expecting start and
> offset or only start, finish.
>
> Is the new API documented yet?
>
> Thanks.
>
> --
> Jonas Bonér
>
> twitter: @jboner
> blog:    http://jonasboner.com
> work:   http://crisp.se
> work:   http://scalablesolutions.se
> code:   http://github.com/jboner
>
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonathan Ellis-3
(either of start and finish may be empty)

On Sat, Aug 8, 2009 at 8:41 AM, Jonathan Ellis<[hidden email]> wrote:

> Stabilizing but not quite finished (329 and 311 are still waiting for
> review).  But the fundamentals are the same.
>
> You have start/finish because that's what defines a range.  You have
> count because you often want the First N results.
>
> -Jonathan
>
> On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:
>> Hey guys.
>>
>> Is the new API stabilizing?
>>
>> How is the new range functions suppose to work?
>>    public List<Column> get_slice_by_names(String keyspace, String
>> key, ColumnParent column_parent, List<byte[]> column_names, int
>> consistency_level) throws InvalidRequestException, NotFoundException,
>> TException;
>>
>>    public List<Column> get_slice(String keyspace, String key,
>> ColumnParent column_parent, byte[] start, byte[] finish, boolean
>> is_ascending, int count, int consistency_level) throws
>> InvalidRequestException, NotFoundException, TException;
>>
>> These both have a start, finish and count. I was expecting start and
>> offset or only start, finish.
>>
>> Is the new API documented yet?
>>
>> Thanks.
>>
>> --
>> Jonas Bonér
>>
>> twitter: @jboner
>> blog:    http://jonasboner.com
>> work:   http://crisp.se
>> work:   http://scalablesolutions.se
>> code:   http://github.com/jboner
>>
>
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonas Bonér-3
2009/8/8 Jonathan Ellis <[hidden email]>:
> (either of start and finish may be empty)

Thanks for fast reply.

So count is the offset from either start or finish
OR
If start and finish is used then count should be -1 ???

>
> On Sat, Aug 8, 2009 at 8:41 AM, Jonathan Ellis<[hidden email]> wrote:
>> Stabilizing but not quite finished (329 and 311 are still waiting for
>> review).  But the fundamentals are the same.
>>
>> You have start/finish because that's what defines a range.  You have
>> count because you often want the First N results.
>>
>> -Jonathan
>>
>> On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:
>>> Hey guys.
>>>
>>> Is the new API stabilizing?
>>>
>>> How is the new range functions suppose to work?
>>>    public List<Column> get_slice_by_names(String keyspace, String
>>> key, ColumnParent column_parent, List<byte[]> column_names, int
>>> consistency_level) throws InvalidRequestException, NotFoundException,
>>> TException;
>>>
>>>    public List<Column> get_slice(String keyspace, String key,
>>> ColumnParent column_parent, byte[] start, byte[] finish, boolean
>>> is_ascending, int count, int consistency_level) throws
>>> InvalidRequestException, NotFoundException, TException;
>>>
>>> These both have a start, finish and count. I was expecting start and
>>> offset or only start, finish.
>>>
>>> Is the new API documented yet?
>>>
>>> Thanks.
>>>
>>> --
>>> Jonas Bonér
>>>
>>> twitter: @jboner
>>> blog:    http://jonasboner.com
>>> work:   http://crisp.se
>>> work:   http://scalablesolutions.se
>>> code:   http://github.com/jboner
>>>
>>
>



--
Jonas Bonér

twitter: @jboner
blog:    http://jonasboner.com
work:   http://crisp.se
work:   http://scalablesolutions.se
code:   http://github.com/jboner
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonathan Ellis-3
Count is always the max number of results to return.

So it means, starting with `start`, or the first one if start is
empty, go until you hit `finish` or `count`, whichever comes first.
Empty is not a legal column name so if finish is empty it is ignored
and only count is used.

We don't offer a numeric offset since that can't be supported
efficiently with a log-structured merge disk format.

On Sat, Aug 8, 2009 at 8:57 AM, Jonas Bonér<[hidden email]> wrote:

> 2009/8/8 Jonathan Ellis <[hidden email]>:
>> (either of start and finish may be empty)
>
> Thanks for fast reply.
>
> So count is the offset from either start or finish
> OR
> If start and finish is used then count should be -1 ???
>
>>
>> On Sat, Aug 8, 2009 at 8:41 AM, Jonathan Ellis<[hidden email]> wrote:
>>> Stabilizing but not quite finished (329 and 311 are still waiting for
>>> review).  But the fundamentals are the same.
>>>
>>> You have start/finish because that's what defines a range.  You have
>>> count because you often want the First N results.
>>>
>>> -Jonathan
>>>
>>> On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:
>>>> Hey guys.
>>>>
>>>> Is the new API stabilizing?
>>>>
>>>> How is the new range functions suppose to work?
>>>>    public List<Column> get_slice_by_names(String keyspace, String
>>>> key, ColumnParent column_parent, List<byte[]> column_names, int
>>>> consistency_level) throws InvalidRequestException, NotFoundException,
>>>> TException;
>>>>
>>>>    public List<Column> get_slice(String keyspace, String key,
>>>> ColumnParent column_parent, byte[] start, byte[] finish, boolean
>>>> is_ascending, int count, int consistency_level) throws
>>>> InvalidRequestException, NotFoundException, TException;
>>>>
>>>> These both have a start, finish and count. I was expecting start and
>>>> offset or only start, finish.
>>>>
>>>> Is the new API documented yet?
>>>>
>>>> Thanks.
>>>>
>>>> --
>>>> Jonas Bonér
>>>>
>>>> twitter: @jboner
>>>> blog:    http://jonasboner.com
>>>> work:   http://crisp.se
>>>> work:   http://scalablesolutions.se
>>>> code:   http://github.com/jboner
>>>>
>>>
>>
>
>
>
> --
> Jonas Bonér
>
> twitter: @jboner
> blog:    http://jonasboner.com
> work:   http://crisp.se
> work:   http://scalablesolutions.se
> code:   http://github.com/jboner
>
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonas Bonér-3
Got it. Thanks a lot for amazing support on a Sat. :-)

2009/8/8 Jonathan Ellis <[hidden email]>:

> Count is always the max number of results to return.
>
> So it means, starting with `start`, or the first one if start is
> empty, go until you hit `finish` or `count`, whichever comes first.
> Empty is not a legal column name so if finish is empty it is ignored
> and only count is used.
>
> We don't offer a numeric offset since that can't be supported
> efficiently with a log-structured merge disk format.
>
> On Sat, Aug 8, 2009 at 8:57 AM, Jonas Bonér<[hidden email]> wrote:
>> 2009/8/8 Jonathan Ellis <[hidden email]>:
>>> (either of start and finish may be empty)
>>
>> Thanks for fast reply.
>>
>> So count is the offset from either start or finish
>> OR
>> If start and finish is used then count should be -1 ???
>>
>>>
>>> On Sat, Aug 8, 2009 at 8:41 AM, Jonathan Ellis<[hidden email]> wrote:
>>>> Stabilizing but not quite finished (329 and 311 are still waiting for
>>>> review).  But the fundamentals are the same.
>>>>
>>>> You have start/finish because that's what defines a range.  You have
>>>> count because you often want the First N results.
>>>>
>>>> -Jonathan
>>>>
>>>> On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:
>>>>> Hey guys.
>>>>>
>>>>> Is the new API stabilizing?
>>>>>
>>>>> How is the new range functions suppose to work?
>>>>>    public List<Column> get_slice_by_names(String keyspace, String
>>>>> key, ColumnParent column_parent, List<byte[]> column_names, int
>>>>> consistency_level) throws InvalidRequestException, NotFoundException,
>>>>> TException;
>>>>>
>>>>>    public List<Column> get_slice(String keyspace, String key,
>>>>> ColumnParent column_parent, byte[] start, byte[] finish, boolean
>>>>> is_ascending, int count, int consistency_level) throws
>>>>> InvalidRequestException, NotFoundException, TException;
>>>>>
>>>>> These both have a start, finish and count. I was expecting start and
>>>>> offset or only start, finish.
>>>>>
>>>>> Is the new API documented yet?
>>>>>
>>>>> Thanks.
>>>>>
>>>>> --
>>>>> Jonas Bonér
>>>>>
>>>>> twitter: @jboner
>>>>> blog:    http://jonasboner.com
>>>>> work:   http://crisp.se
>>>>> work:   http://scalablesolutions.se
>>>>> code:   http://github.com/jboner
>>>>>
>>>>
>>>
>>
>>
>>
>> --
>> Jonas Bonér
>>
>> twitter: @jboner
>> blog:    http://jonasboner.com
>> work:   http://crisp.se
>> work:   http://scalablesolutions.se
>> code:   http://github.com/jboner
>>
>



--
Jonas Bonér

twitter: @jboner
blog:    http://jonasboner.com
work:   http://crisp.se
work:   http://scalablesolutions.se
code:   http://github.com/jboner
Reply | Threaded
Open this post in threaded view
|

Re: new thrift API

Jonas Bonér-3
New API is much better.
Good job.

2009/8/8 Jonas Bonér <[hidden email]>:

> Got it. Thanks a lot for amazing support on a Sat. :-)
>
> 2009/8/8 Jonathan Ellis <[hidden email]>:
>> Count is always the max number of results to return.
>>
>> So it means, starting with `start`, or the first one if start is
>> empty, go until you hit `finish` or `count`, whichever comes first.
>> Empty is not a legal column name so if finish is empty it is ignored
>> and only count is used.
>>
>> We don't offer a numeric offset since that can't be supported
>> efficiently with a log-structured merge disk format.
>>
>> On Sat, Aug 8, 2009 at 8:57 AM, Jonas Bonér<[hidden email]> wrote:
>>> 2009/8/8 Jonathan Ellis <[hidden email]>:
>>>> (either of start and finish may be empty)
>>>
>>> Thanks for fast reply.
>>>
>>> So count is the offset from either start or finish
>>> OR
>>> If start and finish is used then count should be -1 ???
>>>
>>>>
>>>> On Sat, Aug 8, 2009 at 8:41 AM, Jonathan Ellis<[hidden email]> wrote:
>>>>> Stabilizing but not quite finished (329 and 311 are still waiting for
>>>>> review).  But the fundamentals are the same.
>>>>>
>>>>> You have start/finish because that's what defines a range.  You have
>>>>> count because you often want the First N results.
>>>>>
>>>>> -Jonathan
>>>>>
>>>>> On Sat, Aug 8, 2009 at 8:33 AM, Jonas Bonér<[hidden email]> wrote:
>>>>>> Hey guys.
>>>>>>
>>>>>> Is the new API stabilizing?
>>>>>>
>>>>>> How is the new range functions suppose to work?
>>>>>>    public List<Column> get_slice_by_names(String keyspace, String
>>>>>> key, ColumnParent column_parent, List<byte[]> column_names, int
>>>>>> consistency_level) throws InvalidRequestException, NotFoundException,
>>>>>> TException;
>>>>>>
>>>>>>    public List<Column> get_slice(String keyspace, String key,
>>>>>> ColumnParent column_parent, byte[] start, byte[] finish, boolean
>>>>>> is_ascending, int count, int consistency_level) throws
>>>>>> InvalidRequestException, NotFoundException, TException;
>>>>>>
>>>>>> These both have a start, finish and count. I was expecting start and
>>>>>> offset or only start, finish.
>>>>>>
>>>>>> Is the new API documented yet?
>>>>>>
>>>>>> Thanks.
>>>>>>
>>>>>> --
>>>>>> Jonas Bonér
>>>>>>
>>>>>> twitter: @jboner
>>>>>> blog:    http://jonasboner.com
>>>>>> work:   http://crisp.se
>>>>>> work:   http://scalablesolutions.se
>>>>>> code:   http://github.com/jboner
>>>>>>
>>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> Jonas Bonér
>>>
>>> twitter: @jboner
>>> blog:    http://jonasboner.com
>>> work:   http://crisp.se
>>> work:   http://scalablesolutions.se
>>> code:   http://github.com/jboner
>>>
>>
>
>
>
> --
> Jonas Bonér
>
> twitter: @jboner
> blog:    http://jonasboner.com
> work:   http://crisp.se
> work:   http://scalablesolutions.se
> code:   http://github.com/jboner
>



--
Jonas Bonér

twitter: @jboner
blog:    http://jonasboner.com
work:   http://crisp.se
work:   http://scalablesolutions.se
code:   http://github.com/jboner