Skip to content

Commit 9789606

Browse files
committed
HIVE-26830: Update TPCDS30TB metastore dump with histograms
1 parent c18d0df commit 9789606

File tree

95 files changed

+2787
-2977
lines changed

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

95 files changed

+2787
-2977
lines changed
Lines changed: 33 additions & 33 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,10 @@
11
CBO PLAN:
22
HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
33
HiveProject(c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
4-
HiveJoin(condition=[AND(=($3, $7), >($4, $6))], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
5-
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
6-
HiveProject(c_customer_sk=[$0], c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
7-
HiveTableScan(table=[[default, customer]], table:alias=[customer]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
4+
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
5+
HiveProject(c_customer_sk=[$0], c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
6+
HiveTableScan(table=[[default, customer]], table:alias=[customer]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
7+
HiveJoin(condition=[AND(=($1, $5), >($2, $4))], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
88
HiveJoin(condition=[=($3, $1)], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
99
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
1010
HiveFilter(condition=[IS NOT NULL($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
@@ -19,27 +19,27 @@ HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100]): rowcount = ###Masked###, cum
1919
HiveProject(s_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
2020
HiveFilter(condition=[=($24, _UTF-16LE'NM')]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
2121
HiveTableScan(table=[[default, store]], table:alias=[store]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
22-
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
23-
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
24-
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
25-
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
26-
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
27-
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
28-
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
29-
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
30-
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
31-
HiveProject(d_date_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
32-
HiveFilter(condition=[=($6, 2000)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
33-
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
22+
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
23+
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
24+
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
25+
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
26+
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
27+
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[not available]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
28+
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
29+
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
30+
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
31+
HiveProject(d_date_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
32+
HiveFilter(condition=[=($6, 2000)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
33+
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
3434

3535
CBO PLAN:
3636
HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
3737
HiveProject(c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
38-
HiveJoin(condition=[AND(=($3, $7), >($4, $6))], joinType=[inner], algorithm=[none], cost=[{13.140068448808671 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
39-
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[{8.000001214006844E7 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
40-
HiveProject(c_customer_sk=[$0], c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
41-
HiveTableScan(table=[[default, customer]], table:alias=[customer]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
42-
HiveJoin(condition=[=($3, $1)], joinType=[inner], algorithm=[none], cost=[{572.8902844748543 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
38+
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[{8.0000001E7 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
39+
HiveProject(c_customer_sk=[$0], c_customer_id=[$1]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
40+
HiveTableScan(table=[[default, customer]], table:alias=[customer]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
41+
HiveJoin(condition=[AND(=($1, $5), >($2, $4))], joinType=[inner], algorithm=[none], cost=[{13.42239562203678 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
42+
HiveJoin(condition=[=($3, $1)], joinType=[inner], algorithm=[none], cost=[{573.7909187243257 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
4343
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
4444
HiveFilter(condition=[IS NOT NULL($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
4545
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
@@ -53,16 +53,16 @@ HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100]): rowcount = ###Masked###, cum
5353
HiveProject(s_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
5454
HiveFilter(condition=[=($24, _UTF-16LE'NM')]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
5555
HiveTableScan(table=[[default, store]], table:alias=[store]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
56-
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
57-
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
58-
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
59-
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
60-
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
61-
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[{8.180936341080402E9 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
62-
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
63-
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
64-
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
65-
HiveProject(d_date_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
66-
HiveFilter(condition=[=($6, 2000)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
67-
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
56+
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
57+
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
58+
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
59+
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
60+
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
61+
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[{8.180936341080402E9 rows, 0.0 cpu, 0.0 io}]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
62+
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
63+
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
64+
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
65+
HiveProject(d_date_sk=[$0]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
66+
HiveFilter(condition=[=($6, 2000)]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
67+
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim]): rowcount = ###Masked###, cumulative cost = ###Masked###, id = ###Masked###
6868

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,10 @@
11
CBO PLAN:
22
HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100])
33
HiveProject(c_customer_id=[$1])
4-
HiveJoin(condition=[AND(=($3, $7), >($4, $6))], joinType=[inner], algorithm=[none], cost=[not available])
5-
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[not available])
6-
HiveProject(c_customer_sk=[$0], c_customer_id=[$1])
7-
HiveTableScan(table=[[default, customer]], table:alias=[customer])
4+
HiveJoin(condition=[=($2, $0)], joinType=[inner], algorithm=[none], cost=[not available])
5+
HiveProject(c_customer_sk=[$0], c_customer_id=[$1])
6+
HiveTableScan(table=[[default, customer]], table:alias=[customer])
7+
HiveJoin(condition=[AND(=($1, $5), >($2, $4))], joinType=[inner], algorithm=[none], cost=[not available])
88
HiveJoin(condition=[=($3, $1)], joinType=[inner], algorithm=[none], cost=[not available])
99
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2])
1010
HiveFilter(condition=[IS NOT NULL($2)])
@@ -19,16 +19,16 @@ HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100])
1919
HiveProject(s_store_sk=[$0])
2020
HiveFilter(condition=[=($24, _UTF-16LE'NM')])
2121
HiveTableScan(table=[[default, store]], table:alias=[store])
22-
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0])
23-
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))])
24-
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)])
25-
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2])
26-
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)])
27-
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[not available])
28-
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19])
29-
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))])
30-
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns])
31-
HiveProject(d_date_sk=[$0])
32-
HiveFilter(condition=[=($6, 2000)])
33-
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim])
22+
HiveProject(_o__c0=[*(CAST(/($1, $2)):DECIMAL(21, 6), 1.2:DECIMAL(2, 1))], ctr_store_sk=[$0])
23+
HiveFilter(condition=[IS NOT NULL(CAST(/($1, $2)):DECIMAL(21, 6))])
24+
HiveAggregate(group=[{1}], agg#0=[sum($2)], agg#1=[count($2)])
25+
HiveProject(sr_customer_sk=[$0], sr_store_sk=[$1], $f2=[$2])
26+
HiveAggregate(group=[{0, 1}], agg#0=[sum($2)])
27+
HiveJoin(condition=[=($3, $4)], joinType=[inner], algorithm=[none], cost=[not available])
28+
HiveProject(sr_customer_sk=[$2], sr_store_sk=[$6], sr_fee=[$13], sr_returned_date_sk=[$19])
29+
HiveFilter(condition=[AND(IS NOT NULL($6), IS NOT NULL($19))])
30+
HiveTableScan(table=[[default, store_returns]], table:alias=[store_returns])
31+
HiveProject(d_date_sk=[$0])
32+
HiveFilter(condition=[=($6, 2000)])
33+
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim])
3434

Lines changed: 16 additions & 16 deletions
Original file line numberDiff line numberDiff line change
@@ -1,23 +1,23 @@
11
CBO PLAN:
22
HiveSortLimit(sort0=[$0], dir0=[ASC], fetch=[100])
33
HiveProject(i_item_id=[$0], agg1=[/(CAST($1):DOUBLE, $2)], agg2=[CAST(/($3, $4)):DECIMAL(11, 6)], agg3=[CAST(/($5, $6)):DECIMAL(11, 6)], agg4=[CAST(/($7, $8)):DECIMAL(11, 6)])
4-
HiveAggregate(group=[{12}], agg#0=[sum($3)], agg#1=[count($3)], agg#2=[sum($4)], agg#3=[count($4)], agg#4=[sum($6)], agg#5=[count($6)], agg#6=[sum($5)], agg#7=[count($5)])
5-
HiveJoin(condition=[=($1, $11)], joinType=[inner], algorithm=[none], cost=[not available])
6-
HiveJoin(condition=[=($2, $10)], joinType=[inner], algorithm=[none], cost=[not available])
7-
HiveJoin(condition=[=($0, $9)], joinType=[inner], algorithm=[none], cost=[not available])
8-
HiveJoin(condition=[=($7, $8)], joinType=[inner], algorithm=[none], cost=[not available])
4+
HiveAggregate(group=[{1}], agg#0=[sum($5)], agg#1=[count($5)], agg#2=[sum($6)], agg#3=[count($6)], agg#4=[sum($8)], agg#5=[count($8)], agg#6=[sum($7)], agg#7=[count($7)])
5+
HiveJoin(condition=[=($3, $0)], joinType=[inner], algorithm=[none], cost=[not available])
6+
HiveProject(i_item_sk=[$0], i_item_id=[$1])
7+
HiveTableScan(table=[[default, item]], table:alias=[item])
8+
HiveJoin(condition=[=($0, $10)], joinType=[inner], algorithm=[none], cost=[not available])
9+
HiveJoin(condition=[=($7, $9)], joinType=[inner], algorithm=[none], cost=[not available])
10+
HiveJoin(condition=[=($2, $8)], joinType=[inner], algorithm=[none], cost=[not available])
911
HiveProject(cs_bill_cdemo_sk=[$3], cs_item_sk=[$14], cs_promo_sk=[$15], cs_quantity=[$17], cs_list_price=[$19], cs_sales_price=[$20], cs_coupon_amt=[$26], cs_sold_date_sk=[$33])
1012
HiveFilter(condition=[AND(IS NOT NULL($15), IS NOT NULL($3), IS NOT NULL($33))])
1113
HiveTableScan(table=[[default, catalog_sales]], table:alias=[catalog_sales])
12-
HiveProject(d_date_sk=[$0])
13-
HiveFilter(condition=[=($6, 1998)])
14-
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim])
15-
HiveProject(cd_demo_sk=[$0])
16-
HiveFilter(condition=[AND(=($2, _UTF-16LE'W'), =($1, _UTF-16LE'F'), =($3, _UTF-16LE'Primary '))])
17-
HiveTableScan(table=[[default, customer_demographics]], table:alias=[customer_demographics])
18-
HiveProject(p_promo_sk=[$0])
19-
HiveFilter(condition=[OR(=($9, _UTF-16LE'N'), =($14, _UTF-16LE'N'))])
20-
HiveTableScan(table=[[default, promotion]], table:alias=[promotion])
21-
HiveProject(i_item_sk=[$0], i_item_id=[$1])
22-
HiveTableScan(table=[[default, item]], table:alias=[item])
14+
HiveProject(p_promo_sk=[$0])
15+
HiveFilter(condition=[OR(=($9, _UTF-16LE'N'), =($14, _UTF-16LE'N'))])
16+
HiveTableScan(table=[[default, promotion]], table:alias=[promotion])
17+
HiveProject(d_date_sk=[$0])
18+
HiveFilter(condition=[=($6, 1998)])
19+
HiveTableScan(table=[[default, date_dim]], table:alias=[date_dim])
20+
HiveProject(cd_demo_sk=[$0])
21+
HiveFilter(condition=[AND(=($2, _UTF-16LE'W'), =($1, _UTF-16LE'F'), =($3, _UTF-16LE'Primary '))])
22+
HiveTableScan(table=[[default, customer_demographics]], table:alias=[customer_demographics])
2323

0 commit comments

Comments
 (0)