jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-02-02 06:00:36 -05:00
Code Issues Projects Releases Wiki Activity
Files
13bbff4e4e5fc5d459cc6f7a5512f84ceba897cd
impala/testdata/workloads/functional-query/queries/QueryTest/iceberg-is-null-predicate-push-down.test
LPL 7a7934ffba IMPALA-11283: Push-down IS_NULL and NOT_NULL predicates to iceberg 
This patch implements pushing the IS_NULL and NOT_NULL predicates down
to Iceberg.

Testing:
  - Added end-to-end test

Change-Id: I9c3608af67b552bebc55dcc5526f61f5439967bf
Reviewed-on: http://gerrit.cloudera.org:8080/18504
Reviewed-by: Zoltan Borok-Nagy <boroknagyz@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
2022-05-13 23:10:17 +00:00

403 lines
8.5 KiB
Plaintext
Raw Blame History

====
---- QUERY
# Test BIGINT/DOUBLE/STRING/TIMESTAMP/DATE/DECIMAL IsNullPredicate push-down
create table ice_is_null_pred_pd (
col_i INT,
col_bi BIGINT,
col_db DOUBLE,
col_str STRING,
col_ts TIMESTAMP,
col_dt DATE,
col_dc1 DECIMAL(9, 3),
col_dc2 DECIMAL(18, 3),
col_dc3 DECIMAL(38, 3)
) partitioned by spec (col_i) stored as iceberg tblproperties ('write.format.default' = 'parquet');
---- RESULTS
'Table has been created.'
====
---- QUERY
insert into
ice_is_null_pred_pd
values
(0, NULL, 2.71820, '_1700-01-01 00:00:00', '1400-01-01 00:00:00', DATE'1400-01-01', 123.450, 1234567890.120, 1234567890123456789.010),
(1, 12345678901, NULL, 'A1700-01-01 00:00:00', '1400-01-01 01:01:01', DATE'1400-01-01', 123.450, 1234567890.120, 1234567890123456789.010),
(2, 12345678902, 2.71822, NULL, '1400-01-01 02:02:02', DATE'1400-01-01', 123.450, 1234567890.120, 1234567890123456789.010),
(3, 12345678903, 2.71823, '1700-01-01 00:00:00', '1500-01-01 00:00:00', DATE'1500-01-01', 123.450, 1234567890.120, 1234567890123456789.010),
(4, 12345678904, 2.71824, '1700-01-01 00:00:00', NULL, DATE'1500-01-01', 123.450, 1234567890.120, 1234567890123456789.010),
(5, 12345678905, 2.71825, '1700-01-01 00:00:00', '1500-01-01 02:02:02', NULL, 123.450, 1234567890.120, 1234567890123456789.010),
(6, 12345678906, 2.71826, '1700-01-01 00:00:00', '1600-01-01 00:00:00', DATE'1600-01-01', NULL, 1234567890.120, 1234567890123456789.010),
(7, 12345678907, 2.71827, '1700-01-01 00:00:00', '1600-01-01 01:01:01', DATE'1600-01-01', 123.450, NULL, 1234567890123456789.010),
(8, 12345678908, 2.71828, '1700-01-01 00:00:00', '1600-01-01 02:02:02', DATE'1600-01-01', 123.450, 1234567890.120, NULL);
select count(1) from ice_is_null_pred_pd;
---- RESULTS
9
---- TYPES
BIGINT
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 0
aggregation(SUM, NumFileMetadataRead): 9
====
---- QUERY
show files in ice_is_null_pred_pd;
---- RESULTS
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=0/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=1/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=2/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=3/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=4/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=5/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=6/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=7/.*.0.parq','.*',''
row_regex:'$NAMENODE/test-warehouse/$DATABASE.db/ice_is_null_pred_pd/data/col_i=8/.*.0.parq','.*',''
---- TYPES
STRING, STRING, STRING
====
---- QUERY
# Start testing predicate push-down for bigint column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_bi is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_bi as double) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_bi as TIMESTAMP) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
col_bi + 1 is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_bi is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for double column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_db is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_db as decimal(6,5)) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_db is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for string column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_str is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
col_i
from
ice_is_null_pred_pd
where
cast(col_str as TIMESTAMP) is null
order by 1;
---- RESULTS
0
1
2
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_str is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for timestamp column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_ts is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_ts as string) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_ts is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for date column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dt is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_dt as string) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dt is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for decimal(9,3) column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc1 is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_dc1 as double) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc1 is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for decimal(18,3) column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc2 is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_dc2 as double) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc2 is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
---- QUERY
# Start testing predicate push-down for decimal(38,3) column
# The IS_NULL predicate matches 1 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc3 is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 1
====
---- QUERY
# The IS_NULL predicate will not push down
select
count(1)
from
ice_is_null_pred_pd
where
cast(col_dc3 as string) is null;
---- RESULTS
1
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 9
====
---- QUERY
# The NOT_NULL predicate matches 8 row group
select
count(1)
from
ice_is_null_pred_pd
where
col_dc3 is not null;
---- RESULTS
8
---- RUNTIME_PROFILE
aggregation(SUM, NumRowGroups): 8
====
Reference in New Issue View Git Blame Copy Permalink