drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Venki Korukanti <venki.koruka...@gmail.com>
Subject Re: Parquet read optimization
Date Thu, 15 Oct 2015 06:24:16 GMT
Feature was added in the week before the release. We haven't got a chance
to test it in detail. Will try to make it default to true in 1.3.

Thanks
Venki

On Wed, Oct 14, 2015 at 11:16 PM, Tomer Shiran <tshiran@dremio.com> wrote:

> Why is the Parquet read optimization not on by default?
>
> Optimizing Reads of Parquet-Backed Tables
>
> Use the store.hive.optimize_scan_with_native_readers option to optimize
> reads of Parquet-backed external tables from Hive. When set to TRUE, this
> option uses Drill native readers instead of the Hive Serde interface,
> resulting in more performant queries of Parquet-backed external tables.
> (Drill 1.2 and later)
>
> Set the store.hive.optimize_scan_with_native_readers option as described in
> the section, "Planning and Execution Options"
> <http://kristinehahn.github.io/drill/docs/planning-and-execution-options/
> >.
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message