I'm currently using parquet through Spark and repeatedly hit the divide
-by-zero error I submitted the Jira ticket 157 for and the PR for the
fix here: https://github.com/apache/incubator-parquet-mr/pull/102
On 01/28/2015 03:27 PM, Daniel Weeks wrote:
If Cheng is ok with an RC release, could we just publish a new candidate?
Obviously, I want to get an official 1.6.0 out, but I'm not sure if we will
get everything done by the time they want to release.
Thoughts?
-Dan
On Wed, Jan 28, 2015 at 12:23 PM, Ryan Blue <[email protected]> wrote:
On 01/28/2015 11:13 AM, Cheng Lian wrote:
Hi team,
The Spark team is planning to cut an RC release of Spark 1.3.0, roughly
around the end of this month. Currently Spark depends on Parquet
1.6.0rc3, and two issues prevent Spark SQL to push filter predicates
down to Parquet tables (PARQUET-136
<https://issues.apache.org/jira/browse/PARQUET-136>, fixed in master
branch; PARQUET-173 <https://issues.apache.org/jira/browse/PARQUET-173>,
PR available <https://github.com/apache/incubator-parquet-mr/pull/108>).
Would anybody please share the plan for 1.6.0 or 1.6.0rc5? Thanks in
advance!
Best,
Cheng
Hi Cheng,
The 1.6.0 release can happen any time as far as the features are
concerned, but it is our first official Apache release for parquet-mr so we
need to get a few tasks done. These are covered by PARQUET-111 and are
things like adding the rat plugin, making sure all of the files have
correct headers, and going through packaging to ensure NOTICE files are
correct for binary and source artifacts.
rb
--
Ryan Blue
Software Engineer
Cloudera, Inc.
