Thanks for the help and confirmation.
Thanks
Shal
-Original Message-
From: Jonathan Leech [mailto:jonat...@gmail.com]
Sent: Wednesday, July 20, 2016 0:35
To: user@phoenix.apache.org
Subject: Re: HBase checkAndPut Support
I suspect you could do what you want in an hbase coprocessor; howe
It does work fine with the sort merge hint. I was just hoping to utilize a
little extra memory to cache the hashes for the smaller of the two tables
and get better perf from a hash join (although, read below - as you
suggest, that is not the case).
I finally did figure out my config issue. The fol
Use sort merge join if both are large tables
On 20 Jul 2016 8:45 p.m., "Nathan Davis"
wrote:
> OK, thanks Mujtaba and Samarth. I added those properties to my
> hbase-site.xml in the same folder as psql.py and sqlline.py, but still get
> the same MaxServerCacheSizeExceededException when executing
OK, thanks Mujtaba and Samarth. I added those properties to my
hbase-site.xml in the same folder as psql.py and sqlline.py, but still get
the same MaxServerCacheSizeExceededException when executing the simple
join. I am using v4.7 on HBase 1.1 and my query is:
> select e.contact_key
> from event_2
Hi William,
Currently we are not supporting index merge optimizations.
Here is the issue I have raised for the same thing. Would be better to
share details there to show the interest.
https://issues.apache.org/jira/browse/PHOENIX-1801
Can't you combine the columns in both indexes into a single i
Hi all,
I have a question about global secondary index in Phoenix 4.6. See the
following statements:
create table yy (pk integer primary key, cf.a integer, cf.b integer);
create index yya on yy(cf.a);
create index yyb on yy(cf.b);
then upsert some data into table yy; do the following query: