navis requested code review of "HIVE-3617 [jira] Predicates pushed down to hbase is not handled properly when constant part is shown first".
Reviewers: JIRA

DPAL-1922 Predicates pushed down to hbase is not handled properly when constant part is shown first

Test result could not show the difference because predicates pushed down are not removed currently(HIVE-2897). So I added log message(scan.toMap()) and checked the output.

with query
select * from hbase_ppd_keyrange where key > 8 and key < 21;
timeRange=[0, 9223372036854775807], batch=-1, startRow=\x00\x00\x00\x08\x00, stopRow=\x00\x00\x00\x15, ...

but with query
select * from hbase_ppd_keyrange where 8 < key and key < 21;
timeRange=[0, 9223372036854775807], batch=-1, startRow=, stopRow=\x00\x00\x00\x15, ...

TEST PLAN
EMPTY

REVISION DETAIL
https://reviews.facebook.net/D6183

AFFECTED FILES
hbase-handler/src/java/org/apache/hadoop/hive/hbase/HiveHBaseTableInputFormat.java
ql/src/java/org/apache/hadoop/hive/ql/index/IndexPredicateAnalyzer.java
ql/src/java/org/apache/hadoop/hive/ql/index/IndexSearchCondition.java
ql/src/java/org/apache/hadoop/hive/ql/plan/ExprNodeDescUtils.java

MANAGE HERALD DIFFERENTIAL RULES
https://reviews.facebook.net/herald/view/differential/

WHY DID I GET THIS EMAIL?
https://reviews.facebook.net/herald/transcript/14763/

To: JIRA, navis

navis updated the revision "HIVE-3617 [jira] Predicates pushed down to hbase is not handled properly when constant part is shown first".
Reviewers: JIRA

add filp method to genericUDF

REVISION DETAIL
https://reviews.facebook.net/D6183

AFFECTED FILES
ql/src/java/org/apache/hadoop/hive/ql/index/IndexPredicateAnalyzer.java
ql/src/java/org/apache/hadoop/hive/ql/plan/ExprNodeDescUtils.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDF.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFBridge.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPEqual.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPEqualNS.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPEqualOrGreaterThan.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPEqualOrLessThan.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPGreaterThan.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPLessThan.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPNotEqual.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPNotNull.java
ql/src/java/org/apache/hadoop/hive/ql/udf/generic/GenericUDFOPNull.java

To: JIRA, navis

I don't see where you are using newly introduced negate() in this patch? Can that be removed from the patch? Also, patch needs to be rebased on trunk.

Overall: -1 at least one tests failed

Here are the results of testing the latest attachment:
https://issues.apache.org/jira/secure/attachment/12623853/HIVE-3617.3.patch.txt

ERROR: -1 due to 1 failed/errored test(s), 4943 tests executed
Failed tests:

org.apache.hadoop.hive.ql.metadata.TestHiveRemote.testTable

Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/960/testReport
Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/960/console

Messages:

Executing org.apache.hive.ptest.execution.PrepPhase
Executing org.apache.hive.ptest.execution.ExecutionPhase
Executing org.apache.hive.ptest.execution.ReportingPhase
Tests exited with: TestsFailedException: 1 tests failed

This message is automatically generated.

ATTACHMENT ID: 12623853

Patch looks good. I may be confused but won't flip for >= should be <, instead of <=. Similarly, for <, flip should be >= and so on...

I also confused with that. negate() had the semantic as you described (> for <=, etc), which is now removed.

In a word, "3<a" is "a>3", not "a>=3"

Ya, you are correct. Also, some of constant folding code in here won't be needed after HIVE-5771 perhaps we can simplify that whenever thats gets checked in.
+1 lets get this one in.

Committed to trunk. Thanks, Ashutosh.