Re: What kind of policies are used for "Extract Fact Table Distinct Columns" in cube buiding?

[ShaoFeng Shi <sh...@apache.org>]

What kind of policies are used for "Extract Fact Table Distinct Columns" in
cube buiding?

==> The "extract fact distinct" step will fetch the distinct values of each
dimension (if you select to use dictionary as encoding). Usually the output
of this step is much smaller than source table.

My original hive table has 7+ M records, but I got 5+ M records in the
cube. What is the problem?

==> check whether you use "inner join" or have "filter condition" in the
model definition, that will filter some records.


2016-05-24 13:56 GMT+08:00 zhoudg@csdn.net <zh...@csdn.net>:

>
>
> ------------------------------
> 周德高
> zhoudg@csdn.net
>
>
> *发件人：* zhoudg@csdn.net
> *发送时间：* 2016-05-24 13:52
> *收件人：* issues <is...@kylin.apache.org>
> *主题：* What kind of policies are used for "Extract Fact Table Distinct
> Columns" in cube buiding?
> What kind of policies are used for "Extract Fact Table Distinct Columns"
> in cube buiding?
> I found the extracted records are much less than the original records.
> My original hive table has 7+ M records, but I got 5+ M records in the
> cube. What is the problem?
>
> ------------------------------
> 周德高
> zhoudg@csdn.net
>
>


-- 
Best regards,

Shaofeng Shi