You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@kylin.apache.org by Mars J <xu...@gmail.com> on 2016/05/18 06:34:32 UTC

About Kylin Query Errors

Hello All,
      After we build a cube successfully, then we use various sql
statements to query it. In my mind , the sql statement will converted to
hbase query command.

      There are some problem I can't find the answer,please help me to
understand kylin and its principle more accurately.

1. If build a cloned cube succefully, it will get a new htable, then which
one is the exactly htable when using the same query statements ? in the
query page, i can see which cube it is, but then I have build the original
one again later, at this moment, my query use the cloned cube not the
original cube , why it doesn't use the original ?

2. I can't query the fact table using 'select * from fact_table', the error
message is 'Error while executing SQL "select * from tb_fact_new LIMIT
50000": When query with topN, only SUM function is allowed.'

3. What select statement can I use is valid in the query page or use kylin
api, in the other email titled ' SQL  statements supported by kylinl
1.5',hongbin ma has answered that is the subset of sql which calcite
supported.  .I find the select sql  calcite supported is as follow:
select:
      SELECT [ STREAM ] [ ALL | DISTINCT ]
          { * | projectItem [, projectItem ]* }
      FROM tableExpression
      [ WHERE booleanExpression ]
      [ GROUP BY { groupItem [, groupItem ]* } ]
      [ HAVING booleanExpression ]
      [ WINDOW windowName AS windowSpec [, windowName AS windowSpec ]* ]
is that other select grammer is not supported except those ?

Thanks.

Re: About Kylin Query Errors

Posted by Li Yang <li...@apache.org>.
About 1. The choice between 2 clone cubes is undefined. Suggest disable one
of them. It does not make much sense to have 2 identical cube at the same
time.

About 2. Try specify the aggregation function or use a group by. Kylin does
not store raw record, so 'select * from fact' is only processed at Kylin's
best effort.

About 3. You have found the right SQL reference. Let's know any SQL syntax
you desire but missing. There will be enhancement.

On Wed, May 18, 2016 at 2:34 PM, Mars J <xu...@gmail.com> wrote:

> Hello All,
>       After we build a cube successfully, then we use various sql
> statements to query it. In my mind , the sql statement will converted to
> hbase query command.
>
>       There are some problem I can't find the answer,please help me to
> understand kylin and its principle more accurately.
>
> 1. If build a cloned cube succefully, it will get a new htable, then which
> one is the exactly htable when using the same query statements ? in the
> query page, i can see which cube it is, but then I have build the original
> one again later, at this moment, my query use the cloned cube not the
> original cube , why it doesn't use the original ?
>
> 2. I can't query the fact table using 'select * from fact_table', the
> error message is 'Error while executing SQL "select * from tb_fact_new
> LIMIT 50000": When query with topN, only SUM function is allowed.'
>
> 3. What select statement can I use is valid in the query page or use kylin
> api, in the other email titled ' SQL  statements supported by kylinl
> 1.5',hongbin ma has answered that is the subset of sql which calcite
> supported.  .I find the select sql  calcite supported is as follow:
> select:
>       SELECT [ STREAM ] [ ALL | DISTINCT ]
>           { * | projectItem [, projectItem ]* }
>       FROM tableExpression
>       [ WHERE booleanExpression ]
>       [ GROUP BY { groupItem [, groupItem ]* } ]
>       [ HAVING booleanExpression ]
>       [ WINDOW windowName AS windowSpec [, windowName AS windowSpec ]* ]
> is that other select grammer is not supported except those ?
>
> Thanks.
>