Repository: impala
Updated Branches:
refs/heads/master 4ab1b9245 -> ea615d1d8
Analytic mem mgmt cleanup
Following on from IMPALA-7403, clean up some of memory management
in AnalyticEvalNode.
Remove an unnecessary MemPool and avoid reallocating tuples
unnecessarily in Reset(). Also switch to
IMPALA-5031: Remove some undefined behavior of NULL pointers
This was found in ExprTest.LiteralExprs. The rules vioalted are:
1. "reference binding to null pointer of type 'long'". This violates
the standard's [dcl.ref]:
a null reference cannot exist in a well-defined program, because
Make collect_diagnostics executable and minor argument-parsing changes.
collect_diagnostics.py was missing a shebang (#!) at its
beginning as well as the executable bit. Since it's
meant to be runnable standalone, I've added those.
I was able to also simplify the argument handling around
the
Repository: impala
Updated Branches:
refs/heads/master 7002ad3b3 -> 4ab1b9245
Ignore flake8 W503 about breaking before operators
PEP8 now reads that breaking before or after binary operators is fine,
and gives the rationale:
But this can hurt readability in two ways: the operators tend
IMPALA-7347: Ignore numFilesErasureCoded in TestShowCreateTable
This table properties only exist for HDFS tables. To get the test work
on local tables, it needs to be ignored.
Change-Id: Icc8494fb91c4777cee662a97f750486aa8e79a8e
Reviewed-on: http://gerrit.cloudera.org:8080/11192
Reviewed-by:
IMPALA-7419: Fix NullPointerException in SimplifyConditionalsRule
In SimplifyConditionalsRule, when simplifying a 'coalesce' where the
result is another coalesce with some children removed, we create a new
Expr for the new coalesce but previously were not analyzing the new
Expr.
This causes
IMPALA-7335/IMPALA-7418: Add logs to HdfsScanNode to debug the issues
IMPALA-7335 and IMPALA-7418 are failing some builds on jenkins.
However, there is no deterministic method to reproduce them
locally and hence it is difficult to figure out the cause of the
failure. From the existing logs, it
IMPALA-7398: Add logged_in_user alias for effective_user
This patch adds an alias to the effective_user function so that
views created in Hive using the logged_in_user function will work
in Impala.
Example:
CREATE VIEW foo.view1 AS SELECT * FROM foo.table1 WHERE
name=logged_in_user();
Tests:
-