kylin-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From ShaoFeng Shi <shaofeng...@apache.org>
Subject Re: incorrect cardinality
Date Fri, 08 Dec 2017 09:51:37 GMT
Kylin uses HyperLogLog to collect the cardinality of each column, so it is
inaccurate, but the error rate should be acceptable.

Please check whether the column order is different in Kylin and Hive, for
example, col "GENDER" is 4th in Kylin but 5th in Hive. In that case you
need re-sync the table.

2017-12-08 10:06 GMT+08:00 Ge Silas <gosoy@live.cn>:

> Can you please try “Calculate Cardinality” in System tab?
>
> Thanks,
> Silas
>
> On 8 Dec 2017, at 10:03 AM, Shuangyin Ge <gosoy@live.cn> wrote:
>
> I am sorry Sonny.
>
> Please ignore the above response…
>
> Thanks,
> Silas
>
> On 8 Dec 2017, at 9:48 AM, Ge Silas <gosoy@live.cn> wrote:
>
> Hi Sonny,
>
> What was the sampling percentage you used?
>
> Best regards,
> Silas
>
> On 7 Dec 2017, at 6:03 AM, Sonny Heer <sonnyheer@gmail.com> wrote:
>
> We have a table in hive which has a gender column (char(1)).  The group by
> shows the following:
>
>
> M 8946041
> 8 9
> F 14215364
>   215400
>
> Kylin shows:
>
> 10 GENDER char(1) 274693
> Looking at the HiveColumnCardinalityJob code I don't see anything
> obviously wrong.  Any idea why that value is wrong in the UI?
>
> Thanks
>
>
>
>
>


-- 
Best regards,

Shaofeng Shi 史少锋

Mime
View raw message