[
https://issues.apache.org/jira/browse/PHOENIX-16?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13888347#comment-13888347
]
abhishek commented on PHOENIX-16:
---------------------------------
Both workaround queries throw this error: Error: Does not support non-standard
or non-equi join conditions. (state=,code=0)
I have these two tables:
create table if not exists customer ( C_CUSTKEY INTEGER not null primary key,
C_NAME VARCHAR, C_ADDRESS VARCHAR, C_NATIONKEY INTEGER, C_PHONE VARCHAR,
C_ACCTBAL VARCHAR, C_MKTSEGMENT VARCHAR, C_COMMENT VARCHAR, C_EMPTY VARCHAR);
create table if not exists orders (O_ORDERKEY INTEGER not null primary key,
O_CUSTKEY INTEGER not null, O_ORDERSTATUS VARCHAR, O_TOTALPRICE DOUBLE,
O_ORDERDATE VARCHAR, O_ORDERPRIORITY VARCHAR, O_CLERK VARCHAR, O_SHIPPRIORITY
INTEGER, O_COMMENT VARCHAR, O_EMPTY VARCHAR);
And, I tried these two queries:
select * from customer c inner join orders o on c.c_custkey=o.o_custkey and
c.c_custkey<10 where o.o_custkey<100;
select * from customer c inner join orders o on c.c_custkey=o.o_custkey and
c.c_custkey<10 and o.o_custkey<100;
Same error for both.
> Join Query optimization
> -----------------------
>
> Key: PHOENIX-16
> URL: https://issues.apache.org/jira/browse/PHOENIX-16
> Project: Phoenix
> Issue Type: Improvement
> Affects Versions: 3.0.0
> Reporter: abhishek
> Assignee: Maryann Xue
>
> select * from t1 inner join t2 on t1.col1 = t2.col2 where t1.id < 1000 and
> t2.id < 1000
> For above query, phoenix first join the tables t1 and t2, and then apply the
> filter t1.id < 1000 and t2.id < 1000. This is very inefficient. This can be
> easily fixed by pushing the filter operation before the join query.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
