[DISCUSSION]implement delta encoding for numeric type column in SORT_COLUMNS

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[DISCUSSION]implement delta encoding for numeric type column in SORT_COLUMNS

David CaiQiang
Hi all,

    Now we plan to implement delta encoding for the numeric type column in SORT_COLUMNS.
   
    1. use delta encoding to encode the numeric type data

    2. write present metadata to the page header, to record the null value.

    3. improve the compression on no-dictionary string column.
        use RLE to compress the array of the lengths in LV encoding.

    Any thoughts, comments and questions ?

Best Regards
David QiangCai
Best Regards
David Cai
Reply | Threaded
Open this post in threaded view
|

Re: [DISCUSSION]implement delta encoding for numeric type column in SORT_COLUMNS

Jacky Li

> 在 2017年4月5日,下午6:31,QiangCai <[hidden email]> 写道:
>
> Hi all,
>
>    Now we plan to implement delta encoding for the numeric type column in
> SORT_COLUMNS.
>
>    1. use delta encoding to encode the numeric type data
>

I think the adaptive data type conversion still apply here, right?

>    2. write present metadata to the page header, to record the null value.
>
>    3. improve the compression on no-dictionary string column.
>        use RLE to compress the array of the lengths in LV encoding.
>

I think local dictionary is more suitable for this case

>    Any thoughts, comments and questions ?
>
> Best Regards
> David QiangCai
>
>
> --
> View this message in context: http://apache-carbondata-mailing-list-archive.1130556.n5.nabble.com/DISCUSSION-implement-delta-encoding-for-numeric-type-column-in-SORT-COLUMNS-tp10124.html
> Sent from the Apache CarbonData Mailing List archive mailing list archive at Nabble.com.



Reply | Threaded
Open this post in threaded view
|

Re: [DISCUSSION]implement delta encoding for numeric type column in SORT_COLUMNS

Liang Chen
Administrator
In reply to this post by David CaiQiang
Hi David

Thanks for your starting this new feature's discussion.
Can you explain what are the major benefits after doing delta encoding for
the numeric type column.

Regards
Liang

2017-04-05 16:01 GMT+05:30 QiangCai <[hidden email]>:

> Hi all,
>
>     Now we plan to implement delta encoding for the numeric type column in
> SORT_COLUMNS.
>
>     1. use delta encoding to encode the numeric type data
>
>     2. write present metadata to the page header, to record the null value.
>
>     3. improve the compression on no-dictionary string column.
>         use RLE to compress the array of the lengths in LV encoding.
>
>     Any thoughts, comments and questions ?
>
> Best Regards
> David QiangCai
>
>
>
> --
> View this message in context: http://apache-carbondata-
> mailing-list-archive.1130556.n5.nabble.com/DISCUSSION-
> implement-delta-encoding-for-numeric-type-column-in-SORT-
> COLUMNS-tp10124.html
> Sent from the Apache CarbonData Mailing List archive mailing list archive
> at Nabble.com.
>



--
Regards
Liang