jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-01-28 09:03:52 -05:00
Code Issues Projects Releases Wiki Activity
Files
ca5ea4aeab57fb3e8ea16a7ea3e2fb6d188c74cb
impala/testdata/workloads/functional-query/queries/QueryTest/orc-stats.test
stiga-huang 331ff4647d IMPALA-11137: Enable proleptic Gregorian Calendar for Hive 
Since HIVE-22589, Hive still uses Julian Calendar for writing dates
before 1582-10-15, whereas Impala uses proleptic Gregorian Calendar.
This affects the results Impala gets when querying tables written by
Hive. Currently, the Avro and ORC formats of date_tbl are suffering this
issue.

This patch enables proleptic Gregorian Calendar for Hive by default.
It also reverts the two commits of IMPALA-9555 which modifies the tests
to satisfy the inconsistent results.

Tests:
 - Ran CORE tests

Change-Id: I6be9c9720dd352d6821cdaa6c64d35ba20473bc0
Reviewed-on: http://gerrit.cloudera.org:8080/18262
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
2022-02-23 20:09:10 +00:00

864 lines
24 KiB
Plaintext
Raw Blame History

====
---- QUERY
select id, bool_col from functional_orc_def.alltypessmall
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
set explain_level=2;
explain select count(*) from functional_orc_def.alltypessmall where tinyint_col = 10
---- RESULTS: VERIFY_IS_SUBSET
' orc statistics predicates: tinyint_col = CAST(10 AS TINYINT)'
====
---- QUERY
# Test on predicate x < min_val for tinyint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where tinyint_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x <= min_val for tinyint.
select count(*) from functional_orc_def.alltypessmall where tinyint_col <= 0
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x >= max_val for tinyint.
select count(*) from functional_orc_def.alltypessmall where tinyint_col >= 9
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for tinyint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where tinyint_col > 9
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where tinyint_col = -1
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where tinyint_col = 10
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < min_val for smallint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where smallint_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x <= min_val for smallint.
select count(*) from functional_orc_def.alltypessmall where smallint_col <= 0
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x >= max_val for smallint.
select count(*) from functional_orc_def.alltypessmall where smallint_col >= 9
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for smallint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where smallint_col > 9
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where smallint_col = -1
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where smallint_col = 10
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < min_val for int. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where int_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x <= min_val for int.
select count(*) from functional_orc_def.alltypessmall where int_col <= 0
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x >= max_val for int.
select count(*) from functional_orc_def.alltypessmall where int_col >= 9
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for int. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where int_col > 9
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where int_col = -1
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where int_col = 10
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < min_val for bigint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where bigint_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x <= min_val for bigint.
select count(*) from functional_orc_def.alltypessmall where bigint_col <= 0
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x >= max_val for bigint.
select count(*) from functional_orc_def.alltypessmall where bigint_col >= 90
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for bigint. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where bigint_col > 90
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where bigint_col = -1
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test EQ predicate with values outside the range.
select count(*) from functional_orc_def.alltypessmall where bigint_col = 100
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall_bool_sorted where bool_col = true;
---- RESULTS
50
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 50
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall_bool_sorted where bool_col = false;
---- RESULTS
50
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 50
====
---- QUERY
# Test on predicate x < min_val for float. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where float_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < a for float that can't filter out any RowGroups.
select count(*) from functional_orc_def.alltypessmall where float_col < 1
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for float. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where float_col > 9.9
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x > a for float that can't filter out any RowGroups.
select count(*) from functional_orc_def.alltypessmall where float_col > 9
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x < min_val for double. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where double_col < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < a for double that can't filter out any RowGroups.
select count(*) from functional_orc_def.alltypessmall where double_col < 1
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for double. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where double_col > 99
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x > a for double that can't filter out any RowGroups.
select count(*) from functional_orc_def.alltypessmall where double_col > 90
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x < min_val for string. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where string_col < "0"
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x <= min_val for string.
select count(*) from functional_orc_def.alltypessmall where string_col <= "0"
---- RESULTS
12
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x >= max_val for string.
select count(*) from functional_orc_def.alltypessmall where string_col >= "9"
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x > max_val for string. No rows should be read by the ORC reader.
select count(*) from functional_orc_def.alltypessmall where string_col > "9"
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on TIMESTAMP predicates. Currently they are not pushed down (IMPALA-10915).
select count(*) from functional_orc_def.alltypessmall where timestamp_col < "2009-01-01 00:00:00"
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on TIMESTAMP predicates. Currently they are not pushed down (IMPALA-10915).
select count(*) from functional_orc_def.alltypessmall where timestamp_col > "2009-04-03 00:24:00.96"
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test on predicate x < min_val for decimal(9,0).
# Due to ORC-517 not included in the current Hive version (3.1.3000.7.2.12.0-104),
# the ORC files have wrong statistics on d1 column showing that its minimum is 0.
# So we still see RowsRead=5 here.
select count(*) from functional_orc_def.decimal_tbl where d1 < 1234
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d1 < 0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x > max_val for decimal(9,0). No rows should be read by the ORC reader.
# Use 132842.0 instead of 132842 to workaround IMPALA-10916.
select count(*) from functional_orc_def.decimal_tbl where d1 > 132842.0
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x < min_val for decimal(20,10). No rows should be read by the ORC
# reader.
select count(*) from functional_orc_def.decimal_tbl where d3 < 1.23456789
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on predicate x > max_val for decimal(20,10). No rows should be read by the ORC
# reader.
select count(*) from functional_orc_def.decimal_tbl where d3 > 12345.6789
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on EQ predicate with a value out of the range.
select count(*) from functional_orc_def.decimal_tbl where d3 = 1.23456788
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test on EQ predicate with a value out of the range.
select count(*) from functional_orc_def.decimal_tbl where d3 = 12345.6799
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d4 > 0.123456789
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d4 < 0.123456789
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d4 >= 0.12345678
---- RESULTS
5
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d4 >= 0.12345679
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test with inverted predicate
select id, bool_col from functional_orc_def.alltypessmall where -1 > int_col
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Mix with partitioning columns
select count(*) from functional_orc_def.alltypessmall where int_col < 0 and year < 2012
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select id, bool_col from functional_orc_def.alltypessmall where int_col < 3 - 3
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select id, bool_col from functional_orc_def.alltypessmall where int_col < 3 - 3
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Without expr rewrite and thus without constant folding, predicates with const exprs
# won't be pushed down.
set enable_expr_rewrites=0;
select id, bool_col from functional_orc_def.alltypessmall where int_col < 3 - 3
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
select id, bool_col from functional_orc_def.alltypessmall where 5 + 5 < int_col
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Without expr rewrite and thus without constant folding, predicates with const exprs
# won't be pushed down.
set enable_expr_rewrites=0;
select id, bool_col from functional_orc_def.alltypessmall where 5 + 5 < int_col
---- RESULTS
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Query that has an implicit cast to a larger integer type
select count(*) from functional_orc_def.alltypessmall where tinyint_col > 1000000000000
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Predicates with explicit casts are not supported when evaluating orc statistics.
select count(*) from functional_orc_def.alltypessmall where '0' > cast(tinyint_col as string)
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Explicit casts between numerical types can violate the transitivity of "min()", so they
# are not supported when evaluating orc statistics.
select count(*) from functional_orc_def.alltypes where cast(id as tinyint) < 10;
---- RESULTS
3878
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 7300
====
---- QUERY
# Predicates on array position can't be pushed down into the orc reader.
select count(*) from functional_orc_def.complextypestbl.int_array where pos < 5;
---- RESULTS
9
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 8
====
---- QUERY
# Test the conversion of constant IN lists to min/max predicats
select count(*) from functional_orc_def.alltypes where int_col in (-1,-2,-3,-4);
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypes where id IN (1,25,49);
---- RESULTS
3
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 310
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col < '0001-06-21';
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col <= '0001-06-21';
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 10
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col > '2018-12-31';
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col >= '2018-12-31';
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 3
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col = '2018-12-31';
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 3
====
---- QUERY
# Test predicates on CHAR type. They are not pushed down (IMPALA-10882). Just make sure
# we don't hit DCHECKs.
select count(*) from functional_orc_def.chars_tiny where cs < cast('1aaaa' as char(5));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 9
====
---- QUERY
# Test predicates on CHAR type. They are not pushed down (IMPALA-10882). Just make sure
# we don't hit DCHECKs.
select count(*) from functional_orc_def.chars_tiny where cs > cast('a' as char(5));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 9
====
---- QUERY
# Test predicates on VARCHAR type. They are not pushed down (IMPALA-10882). Just make sure
# we don't hit DCHECKs.
select count(*) from functional_orc_def.chars_tiny where vc < cast('1cccc' as varchar(32));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 9
====
---- QUERY
# Test predicates on VARCHAR type. They are not pushed down (IMPALA-10882). Just make sure
# we don't hit DCHECKs.
select count(*) from functional_orc_def.chars_tiny where vc > cast('c' as varchar(32));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 9
====
---- QUERY
# Test that stats support can be disabled using the orc_read_statistics query option.
set orc_read_statistics=0;
select count(*) from functional_orc_def.alltypes where id < 0;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 7300
====
---- QUERY
# Test on a larger ORC file that has multiple stripes and each stripe has multiple row
# groups.
select count(*) from tpch_orc_def.lineitem where l_orderkey = 1609411;
---- RESULTS
7
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 13501
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range.
select count(*) from functional_orc_def.alltypessmall where tinyint_col in (-1, 10);
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
select count(*) from functional_orc_def.alltypessmall where tinyint_col in (1, 5);
---- RESULTS
20
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range.
select count(*) from functional_orc_def.alltypessmall where smallint_col in (-1, 10);
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
select count(*) from functional_orc_def.alltypessmall where smallint_col in (1, 5);
---- RESULTS
20
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range.
select count(*) from functional_orc_def.alltypessmall where int_col in (-1, 10);
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
select count(*) from functional_orc_def.alltypessmall where int_col in (1, 5);
---- RESULTS
20
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
# The table has 4 files. 3 of them will be skipped.
select count(*) from functional_orc_def.alltypessmall where id in (1, 3, 5);
---- RESULTS
3
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 25
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range.
select count(*) from functional_orc_def.alltypessmall where bigint_col in (-1, 91);
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
select count(*) from functional_orc_def.alltypessmall where bigint_col in (10, 50);
---- RESULTS
20
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall_bool_sorted where bool_col in (true);
---- RESULTS
50
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 50
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall_bool_sorted where bool_col in (false);
---- RESULTS
50
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 50
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range.
# The range of string_col is ['0', '9']. '/' is smaller than '0' and 'A' is larger
# than '9'.
select count(*) from functional_orc_def.alltypessmall where string_col in ('/', 'A');
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
# Test pushing down IN-list predicate with literals inside the value range.
select count(*) from functional_orc_def.alltypessmall where string_col in ('1', '5');
---- RESULTS
20
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 100
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col in ('0001-06-18', '2019-12-31');
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.date_tbl
where date_part in ("2017-11-27", "1399-06-27") and date_col in ('0001-06-21', '2018-12-31');
---- RESULTS
2
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 13
====
---- QUERY
# Test pushing down IN-list predicate with literals outside the value range. Explicitly cast
# the literals so FE won't wrap 'd1' as casting to INT, so the predicate can be pushed down.
# Due to ORC-517 not included in the current Hive version (3.1.3000.7.2.12.0-104),
# the ORC files have wrong statistics on d1 column showing that its minimum is 0. So here we
# use -1 to be smaller than it. The max of d1 is 132842 so we use 132843.
select count(*) from functional_orc_def.decimal_tbl
where d1 in (cast(-1 as decimal(9,0)), cast(132843 as decimal(9,0)));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl
where d1 in (cast(0 as decimal(9,0)), cast(132842 as decimal(9,0)));
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl
where d1 in (cast(1234 as decimal(9,0)), cast(2345 as decimal(9,0)), cast(12345 as decimal(9,0)));
---- RESULTS
4
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl
where d3 in (cast(-1 as decimal(20,10)), cast(12346 as decimal(20,10)));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl
where d3 in (cast(2 as decimal(20,10)), cast(200 as decimal(20,10)));
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl
where d3 in (cast(12.3456789 as decimal(20,10)), cast(12345.6789 as decimal(20,10)));
---- RESULTS
2
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 5
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where bool_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where tinyint_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where smallint_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where int_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where bigint_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where float_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where double_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where string_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.alltypessmall where timestamp_col is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d1 is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
---- QUERY
select count(*) from functional_orc_def.decimal_tbl where d4 is null;
---- RESULTS
0
---- RUNTIME_PROFILE
aggregation(SUM, RowsRead): 0
====
Reference in New Issue View Git Blame Copy Permalink