kylin-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Luke Han <luke...@gmail.com>
Subject Re: missing a field in the SQL when creating cube
Date Wed, 06 Jan 2016 00:16:55 GMT
it is by design, hierarchy/derived dimension will not be included in
generated SQL which read data from Hive, which associate with PK/FK.

Thanks.




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

Luke Han

On Wed, Jan 6, 2016 at 3:07 AM, Zhang, Zhong <zzhang@cardlytics.com> wrote:

> Hi All,
>
>
>
> I’ve encountered a kind of weird problem. I just normally created a cube
> step by step.
>
> Based on my understanding, after filling all the information in cube info,
> data model,
>
> dimensions, and etc, a SQL will be generated in the SQL column. The
> following is the
>
> generated SQL:
>
>
>
> SELECT
>
> ...
>
> SPENDTRIPINC.POSTALCD
>
> ,FLATGEO2.TOWN
>
> ,FLATGEO2.DMA_CODE
>
> ,FLATGEO2.PROVINCECD
>
> ,FLATGEO2.REGIONNAME
>
> ,FLATGEO2.MSA
>
> ...
>
> FROM FACT.SPENDTRIPINC as SPENDTRIPINC
>
> INNER JOIN FACT.CALENDARDATES as CALENDARDATES
>
> ON SPENDTRIPINC.DIMDATEKEY = CALENDARDATES.CALENDARDAYID
>
> INNER JOIN FACT.FLATGEO2 as FLATGEO2
>
> ON SPENDTRIPINC.POSTALCD = FLATGEO2.POSTALCD
>
>
>
> I do put fields "POSTALCD","TOWN","DMA_CODE","PROVINCECD","REGIONNAME"
> and "MSA"
>
> from table FLATGEO2 as a hierarchy dimension. There should be a filed
> POSTALCD
>
> in the SQL. But unfortunately it is not there just as you see.
>
>
>
> The weird thing is that if checking at JSON(cube) column, POSTALCD is
> there:
>
> {
>
>       "id": 12,
>
>       "name": "FACTGEO",
>
>       "table": "FACT.FLATGEO2",
>
>       "column": [
>
>         "POSTALCD",
>
>         "TOWN",
>
>         "DMA_CODE",
>
>         "PROVINCECD",
>
>         "REGIONNAME",
>
>         "MSA"
>
>       ],
>
>       "derived": null,
>
>       "hierarchy": true
>
>     }
>
>
>
> Best regards,
>
> Zhong
>
>
>

Mime
View raw message