drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Julian Hyde <julianh...@gmail.com>
Subject Another pair of eyes...
Date Tue, 07 May 2013 05:06:08 GMT
Hi Drillers,

I am having problems integrating the logical plan reference interpreter with the Optiq plans
that I generate to implement SQL. Specifically with which "wrapper" elements I should expect
to wrap the rows that come out of the reference interpreter.

I'd appreciate if another Drill developer could run my code in a debugger and see whether
I am generating the wrong logical plan, or interpreting its results wrongly, or whether the
reference interpreter is broken. In other words, I'd like to borrow another pair of eyes.
Any volunteers?

Download the "optiq" branch from my github fork:

$ git clone git@github.com:julianhyde/incubator-drill.git
$ cd incubator-drill
$ git checkout optiq

Then run the "testCount" method of sandbox/prototype/sqlparser/src/test/java/org/apache/drill/jdbc/test/JdbcTest.java.
It ought to return something like this:

C=1; DEPTID=null
C=2; DEPTID=34
C=2; DEPTID=33
C=1; DEPTID=31

but actually returns this:

C=1; DEPTID={DEPTID=null, LASTNAME=John}
C=1; DEPTID={DEPTID=34, LASTNAME=Robinson}
C=1; DEPTID={DEPTID=33, LASTNAME=Steinburg}
C=1; DEPTID={DEPTID=31, LASTNAME=Rafferty}
C=1; DEPTID={DEPTID=33, LASTNAME=Jones}
C=1; DEPTID={DEPTID=34, LASTNAME=Smith}

As you can see, the DEPTID column has an extra level of nesting. I print the logical plan
to stdout, so see whether that makes sense.

Julian
Mime
View raw message