[DRILL-4665] Partition pruning not working for hive partitioned table with 'LIKE' and '=' filter

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.6.0
• Fix Version/s: 1.8.0
• Component/s: Query Planning & Optimization, Storage - Hive
• Assignee: Gautam Parai
• Reporter: Chun Chang
• Created: 10/May/16 21:23
• Updated: 18/Jul/16 17:58
• Resolved: 15/Jul/16 00:00

Description

This problem was initially reported by Shankar Mane <[e-mail]>

> Problem:
>
> 1. In drill, we are using hive partition table. But explain plan (same
> query) for like and = operator differs and used all partitions in case of
> like operator.
> 2. If you see below drill explain plans: Like operator uses all
> partitions where
> = operator uses only partition filtered by log_date condition.

I reproduced the reported issue. I have a partitioned hive external table with the following schema:

create external table if not exists lineitem_partitioned (
    l_orderkey bigint,
    l_partkey bigint,
    l_suppkey bigint,
    l_linenumber bigint,
    l_quantity double,
    l_extendedprice double,
    l_discount double,
    l_tax double,
    l_returnflag string,
    l_linestatus string,
    l_shipdate string,
    l_commitdate string,
    l_receiptdate string,
    l_shipinstruct string,
    l_shipmode string,
    l_comment string
)
partitioned by (year int, month int, day int)
STORED AS PARQUET
LOCATION '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem';

alter table lineitem_partitioned add partition(year=2015, month=1, day=1) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/1';
alter table lineitem_partitioned add partition(year=2015, month=1, day=2) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/2';
alter table lineitem_partitioned add partition(year=2015, month=1, day=3) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/3';
alter table lineitem_partitioned add partition(year=2015, month=1, day=4) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/4';
alter table lineitem_partitioned add partition(year=2015, month=1, day=5) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/5';
alter table lineitem_partitioned add partition(year=2015, month=1, day=6) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/6';
alter table lineitem_partitioned add partition(year=2015, month=1, day=7) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/7';
alter table lineitem_partitioned add partition(year=2015, month=1, day=8) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/8';
alter table lineitem_partitioned add partition(year=2015, month=1, day=9) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/9';
alter table lineitem_partitioned add partition(year=2015, month=1, day=10) location  '/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/10';

Without 'LIKE', drill plans right:

0: jdbc:drill:schema=dfs.drillTestDir> explain plan for select l_shipdate, l_linestatus, l_shipinstruct, `day` from lineitem_partitioned_db2k_2_999 where `day` = 2 limit 10;
+------+------+
| text | json |
+------+------+
| 00-00    Screen
00-01      Project(l_shipdate=[$0], l_linestatus=[$1], l_shipinstruct=[$2], day=[$3])
00-02        SelectionVectorRemover
00-03          Limit(fetch=[10])
00-04            Limit(fetch=[10])
00-05              Project(l_shipdate=[$1], l_linestatus=[$0], l_shipinstruct=[$2], day=[$3])
00-06                Scan(groupscan=[HiveScan [table=Table(dbName:md815_db2k, tableName:lineitem_partitioned_db2k_2_999), columns=[`l_linestatus`, `l_shipdate`, `l_shipinstruct`, `day`], numPartitions=1, partitions= [Partition(values:[2015, 1, 2])], inputDirectories=[maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/2]]])

With 'LIKE', pruning is not happening:

0: jdbc:drill:schema=dfs.drillTestDir> explain plan for select l_shipdate, l_linestatus, l_shipinstruct, `day` from lineitem_partitioned_db2k_2_999 where `day` = 2 and l_shipinstruct like '%BACK%' limit 10;
+------+------+
| text | json |
+------+------+
| 00-00    Screen
00-01      Project(l_shipdate=[$0], l_linestatus=[$1], l_shipinstruct=[$2], day=[$3])
00-02        SelectionVectorRemover
00-03          Limit(fetch=[10])
00-04            Limit(fetch=[10])
00-05              Project(l_shipdate=[$1], l_linestatus=[$0], l_shipinstruct=[$2], day=[$3])
00-06                SelectionVectorRemover
00-07                  Filter(condition=[AND(=($3, 2), LIKE($2, '%BACK%'))])
00-08                    Scan(groupscan=[HiveScan [table=Table(dbName:md815_db2k, tableName:lineitem_partitioned_db2k_2_999), columns=[`l_linestatus`, `l_shipdate`, `l_shipinstruct`, `day`], numPartitions=10, partitions= [Partition(values:[2015, 1, 1]), Partition(values:[2015, 1, 10]), Partition(values:[2015, 1, 2]), Partition(values:[2015, 1, 3]), Partition(values:[2015, 1, 4]), Partition(values:[2015, 1, 5]), Partition(values:[2015, 1, 6]), Partition(values:[2015, 1, 7]), Partition(values:[2015, 1, 8]), Partition(values:[2015, 1, 9])], inputDirectories=[maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/1, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/10, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/2, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/3, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/4, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/5, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/6, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/7, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/8, maprfs:/drill/testdata/tpch100_dir_partitioned_50000files/lineitem/2015/1/9]]])

Issue Links

• links to: GitHub Pull Request #526