phoenix-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Samarth Jain <sama...@apache.org>
Subject Re: HBase rowkey filter impl in Phoenix for scanning specific time range rows
Date Tue, 25 Aug 2015 00:11:51 GMT
Hi Sun,

There is no custom HBase filter that phoenix uses to scan specific time
range rows.

Having said that, I am currently working on
https://issues.apache.org/jira/browse/PHOENIX-914 that is going to provide
the capability of having a column directly map to HBase cell level
timestamp. By specifying the time component of your row key as this special
column type, it would help Phoenix use an HBase level optimization that
skips HFiles to scan on the basis of time ranges (min and max cell
timestamps) those HFiles store. This optimization should help for typical
use cases that we see for time series based data where the time component
(event_date, created_date, modified_date, etc.) could be naturally mapped
to the HBase cell timestamp.

Stay tuned :)

- Samarth



On Mon, Aug 24, 2015 at 12:56 AM, Fulin Sun <sunfl@certusnet.com.cn> wrote:

> Hi, team,
>
> Being interested in whether Phoenix has such custom hbase filter to scan
> specific time range rows. Our use case is that
>
> timestamp being part of the rowkey, and we need to filter rows according
> to time range. Would be good for you experts
>
> can lead me to the proper implementation of Phoenix source code.
>
> Best,
> Sun.
>
> ------------------------------
> ------------------------------
>
> *Certus**Net*
>
>
>

Mime
View raw message