Re: Kylin cube partitioning strategy

[Luke Han <lu...@gmail.com>]

Hi Norbert,
     Typically, people are using BI tool rather than SQL client to
interactive with Kylin.
     But I think this is valid feature to make popular SQL tools also works
with Kylin.

     Forward this mail to dev@ list for more comments.

     Thanks.
Luke


Best Regards!
---------------------

Luke Han

On Tue, Mar 15, 2016 at 10:10 PM, Norbert Burger <no...@gmail.com>
wrote:

> Thanks Li and Dong for the quick response.  Although this issue is not
> critical, it makes Workbench with Kylin mostly unusable, since every query
> creates the Messages panel and moves focus there.
>
> There is a similar Hive JIRA open here:
> https://issues.apache.org/jira/browse/HIVE-7680
>
> I'll take a look at the code, do you know if this bug is on the Kylin side
> or the Calcite side?
>
> Until the driver is fixed, any there any team recommendations on a SQL GUI
> which doesn't assume getMoreResults() is implemented?
>
> Norbert
>
> On Tue, Mar 15, 2016 at 2:51 AM, Dong Li <li...@apache.org> wrote:
>
>> Hi Norbert,
>>
>>
>> I tried SQL Workbench with Kylin jdbc driver, and found that even though
>> NPE thrown from the caller(WorkBench), we still can get query result in
>> “Result" panel. Hoping it’s not blocking?
>>
>> For this issue, let's track it on
>> https://issues.apache.org/jira/browse/KYLIN-1494
>>
>> Thanks,
>> Dong Li
>>
>>  Original Message
>> *Sender:* Li Yang<li...@apache.org>
>> *Recipient:* user<us...@kylin.apache.org>
>> *Date:* Tuesday, Mar 15, 2016 16:02
>> *Subject:* Re: Kylin cube partitioning strategy
>>
>> > Does Kylin support an hourly partitioning strategy (yyyy-MM-dd-hh)?
>> One of our incoming datasets with which I want to create a cube is
>> partitioned at the hourly level.  If not, how do folks handle this grain
>> difference?
>>
>> The "date format" of the partition date column can be set to include the
>> hour. The "date format" may not be editable in the GUI depending on the
>> Kylin version you are using. What's your version?
>>
>> > SQL Workbench seems not to play well with the Kylin JDBC driver (every
>> query throws an NPE at workbench.sql.SqlCommand.processResults).  Has
>> anyone had luck with this setup?
>>
>> This deserves a JIRA. https://issues.apache.org/jira/browse/KYLIN-1494
>>
>> On Tue, Mar 15, 2016 at 5:12 AM, Norbert Burger <norbert.burger@gmail.com
>> > wrote:
>>
>>> Hi folks,
>>>
>>> * Does Kylin support an hourly partitioning strategy (yyyy-MM-dd-hh)?
>>> One of our incoming datasets with which I want to create a cube is
>>> partitioned at the hourly level.  If not, how do folks handle this grain
>>> difference?
>>>
>>> * SQL Workbench seems not to play well with the Kylin JDBC driver (every
>>> query throws an NPE at workbench.sql.SqlCommand.processResults).  Has
>>> anyone had luck with this setup?
>>>
>>> Norbert
>>>
>>>
>>
>