Files
impala/tests/query_test/test_expr_limits.py
Alex Behm 1d460474af IMPALA-432: Impose a safe maximum expr-tree depth and a safe maximum number of expr children.
Change-Id: Ib519ffd5cd069b676850598b0b30b50b368cb23b
Reviewed-on: http://gerrit.ent.cloudera.com:8080/692
Reviewed-by: Marcel Kornacker <marcel@cloudera.com>
Tested-by: jenkins
Reviewed-by: Alex Behm <alex.behm@cloudera.com>
Tested-by: Alex Behm <alex.behm@cloudera.com>
2014-01-08 10:53:28 -08:00

95 lines
3.4 KiB
Python

#!/usr/bin/env python
# Copyright (c) 2012 Cloudera, Inc. All rights reserved.
# Validates limit on scan nodes
#
import logging
import pytest
from copy import copy
from tests.common.impala_test_suite import ImpalaTestSuite
from tests.common.test_vector import *
from tests.common.test_dimensions import create_exec_option_dimension
# Tests very deep expression trees and expressions with many children. Impala defines
# a 'safe' upper bound on the expr depth and the number of expr children in the
# FE Expr.java and any changes to those limits should be reflected in this test.
# The expr limits primarily guard against stack overflows or similar problems
# causing crashes. Therefore, this tests succeeds if no Impalads crash.
class TestExprLimits(ImpalaTestSuite):
# Keep these in sync with Expr.java
EXPR_CHILDREN_LIMIT = 10000;
EXPR_DEPTH_LIMIT = 2000;
@classmethod
def get_workload(self):
return 'functional-query'
@classmethod
def add_test_dimensions(cls):
super(TestExprLimits, cls).add_test_dimensions()
cls.TestMatrix.clear_dimension('exec_option')
# Run with codegen enabled and disabled.
cls.TestMatrix.add_dimension(create_exec_option_dimension(
cluster_sizes=[0], disable_codegen_options=[False, True], batch_sizes=[0]))
# There is no reason to run these tests using all dimensions.
cls.TestMatrix.add_constraint(lambda v:\
v.get_value('table_format').file_format == 'text' and\
v.get_value('table_format').compression_codec == 'none')
def test_expr_child_limit(self, vector):
# IN predicate
in_query = "select 1 IN("
for i in xrange(0, self.EXPR_CHILDREN_LIMIT - 1):
in_query += str(i)
if (i + 1 != self.EXPR_CHILDREN_LIMIT - 1):
in_query += ","
in_query += ")"
self.__exec_query(in_query)
# CASE expr
case_query = "select case "
for i in xrange(0, self.EXPR_CHILDREN_LIMIT/2):
case_query += " when true then 1"
case_query += " end"
self.__exec_query(case_query)
def test_expr_depth_limit(self, vector):
# Compound predicates
and_query = "select " + self.__gen_deep_infix_expr("true", " and false")
self.__exec_query(and_query)
or_query = "select " + self.__gen_deep_infix_expr("true", " or false")
self.__exec_query(or_query)
# Arithmetic expr
arith_query = "select " + self.__gen_deep_infix_expr("1", " + 1")
self.__exec_query(arith_query)
# TODO: Test function-call exprs after IMPALA-621 has been fixed.
#func_query = "select " + self.__gen_deep_func_expr("lower(", "'abc'", ")")
#self.__exec_query(func_query)
# Casts.
cast_query = "select " + self.__gen_deep_func_expr("cast(", "1", " as int)")
self.__exec_query(cast_query)
def __gen_deep_infix_expr(self, prefix, repeat_suffix):
expr = prefix
for i in xrange(self.EXPR_DEPTH_LIMIT - 1):
expr += repeat_suffix
return expr
def __gen_deep_func_expr(self, open_func, base_arg, close_func):
expr = ""
for i in xrange(self.EXPR_DEPTH_LIMIT - 1):
expr += open_func
expr += base_arg
for i in xrange(self.EXPR_DEPTH_LIMIT - 1):
expr += close_func
return expr
def __exec_query(self, sql_str):
try:
impala_ret = self.execute_query(sql_str)
assert impala_ret.success, "Failed to execute query %s" % (sql_str)
except: # consider any exception a failure
assert False, "Failed to execute query %s" % (sql_str)