A join between bigint and string might actually be evaluated by converting both
values to a double .. try doing an EXPLAIN of the query, that might show what
conversion is being done for the join keys.
If that is the case, you could try explicitly casting the join keys to either
string or bigin
No, The join columns are ("bigint" and "string").
Yong
Subject: RE: Hive 0.12 MAPJOIN hangs sometimes
From: pjayachand...@hortonworks.com
To: user@hive.apache.org; user@hive.apache.org
Date: Fri, 11 Mar 2016 15:48:52 +
Is the join column of type double? If so there is a known issue wit
I believe this is an issue in the Storm Hive bolt. I don’t have an Apache JIRA
on it, but if you ask on the Hortonworks lists we can connect you with the fix
for the storm bolt.
Alan.
> On Mar 10, 2016, at 04:02, Igor Kuzmenko wrote:
>
> Hello, I'm using Hortonworks Data Platform 2.3.4 which
Is the join column of type double? If so there is a known issue with
DoubleWritable hash collisions that makes hash join insanely slow.
Thanks
Prasanth
On Fri, Mar 11, 2016 at 7:33 AM -0800, "Yong Zhang"
mailto:java8...@hotmail.com>> wrote:
I understand the Hive version problem.
We are usi
I understand the Hive version problem.
We are using IBM BigInsights V3.0.0.2, which comes with Hadoop 2.2.0 and Hive
0.12. It is extremely difficult to upgrade to BigInsights v4.x, as IBM did V4
totally different as V3. We are looking for the option to upgrade, but it won't
be a fast way.
The q
Honestly 0.12 is a no go - you miss a lot of performance improvements. Probably
your query would execute in less than a minute. If your Hadoop vendor does not
support smooth upgrades then change it. Hive 1.2.1 is the absolute minimum
including using Orc or parquet as a table format and tez (pref
Hi, Hive users:
Currently our Hadoop vendor comes with Hive 0.12. I know it is a kind of old
version, but upgrade still has some long path to go.
Right now, we are facing an issue in the Hive 0.12.
We have one ETL kind of steps implemented in Hive, and due to the data volume
in this step, we know
This one seems related
https://issues.apache.org/jira/browse/YARN-4538
Yet to ascertain if it actually fixes this issue.
On Thu, Mar 10, 2016 at 11:43 PM, Gopal Vijayaraghavan
wrote:
>
> > This seems to be something YARN fair-scheduler reporting it this way..
> >although Tez doesn't seem to ha
yes you can
Register table in hive based in hbase
CREATE EXTERNAL TABLE IF NOT EXISTS DB_NAME.TABLE_NAME(COL1 STRING,COL2
STRING,COL3 STRING)
STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler'
WITH SERDEPROPERTIES ("hbase.columns.mapping" = ":key,cf:COL2,cf:COL3")
TBLPROPERTIES ("hbase.t
Hi,
I have a HBase table with rowkey and column family.
Is it possible to map HBase table to hive table?
Thanks
10 matches
Mail list logo
