jprdonnelly/impala
1
0
Fork 0
mirror of https://github.com/apache/impala.git synced 2026-01-08 12:02:54 -05:00
Code Issues Projects Releases Wiki Activity
Files
e94de02469a3dfeda7d9358d7a78cb54c7a67159
impala/testdata/workloads/functional-planner/queries/PlannerTest/tpch-all.test
ishaan 2b5df0c6ff [CDH5] Convert tpch schemas to decimal and change the queries where possible. 
I used the following document for reference: http://www.tpc.org/tpch/spec/tpch2.1.0.pdf

Change-Id: Ic84db0628323c90e89552707f214bbb9fa2f2ae0
Reviewed-on: http://gerrit.ent.cloudera.com:8080/3132
Reviewed-by: Ishaan Joshi <ishaan@cloudera.com>
Tested-by: jenkins
2014-07-08 14:51:43 -07:00

2246 lines
62 KiB
Plaintext
Raw Blame History

# TODO: Change these tests to use the partitioned TPCH tables.
# Q1 - Pricing Summary Report Query
# Modifications: Remove ORDER BY, added ROUND() calls
select
l_returnflag,
l_linestatus,
round(sum(l_quantity), 1),
round(sum(l_extendedprice), 1),
round(sum(l_extendedprice * (1 - l_discount)), 1),
round(sum(l_extendedprice * (1 - l_discount) * (1 + l_tax)), 1),
round(avg(l_quantity), 1),
round(avg(l_extendedprice), 1),
round(avg(l_discount), 1), count(1)
from
tpch.lineitem
where
l_shipdate<='1998-09-02'
group by
l_returnflag,
l_linestatus
---- PLAN
01:AGGREGATE [FINALIZE]
| output: sum(l_quantity), sum(l_extendedprice), sum(l_extendedprice * (1 - l_discount)), sum(l_extendedprice * (1 - l_discount) * (1 + l_tax)), count(l_quantity), count(l_extendedprice), sum(l_discount), count(l_discount), count(1)
| group by: l_returnflag, l_linestatus
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate <= '1998-09-02'
---- DISTRIBUTEDPLAN
04:EXCHANGE [UNPARTITIONED]
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_quantity)), sum(sum(l_extendedprice)), sum(sum(l_extendedprice * (1 - l_discount))), sum(sum(l_extendedprice * (1 - l_discount) * (1 + l_tax))), sum(count(l_quantity)), sum(count(l_extendedprice)), sum(sum(l_discount)), sum(count(l_discount)), sum(count(1))
| group by: l_returnflag, l_linestatus
|
02:EXCHANGE [HASH(l_returnflag,l_linestatus)]
|
01:AGGREGATE
| output: sum(l_quantity), sum(l_extendedprice), sum(l_extendedprice * (1 - l_discount)), sum(l_extendedprice * (1 - l_discount) * (1 + l_tax)), count(l_quantity), count(l_extendedprice), sum(l_discount), count(l_discount), count(1)
| group by: l_returnflag, l_linestatus
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate <= '1998-09-02'
====
# Q2 - Minimum Cost Supplier Query
insert overwrite table tpch.q2_minimum_cost_supplier_tmp1
select
s.s_acctbal,
s.s_name,
n.n_name,
p.p_partkey,
ps.ps_supplycost,
p.p_mfgr,
s.s_address,
s.s_phone,
s.s_comment
from tpch.partsupp ps
join tpch.part p
on (p.p_partkey = ps.ps_partkey and p.p_size = 15 and p.p_type like '%BRASS')
join tpch.supplier s
on (s.s_suppkey = ps.ps_suppkey)
join tpch.nation n
on (s.s_nationkey = n.n_nationkey)
join tpch.region r
on (n.n_regionkey = r.r_regionkey and r.r_name = 'EUROPE')
---- PLAN
WRITE TO HDFS [tpch.q2_minimum_cost_supplier_tmp1, OVERWRITE=true]
| partitions=1
|
08:HASH JOIN [INNER JOIN]
| hash predicates: n.n_regionkey = r.r_regionkey
|
|--04:SCAN HDFS [tpch.region r]
| partitions=1/1 size=384B compact
| predicates: r.r_name = 'EUROPE'
|
07:HASH JOIN [INNER JOIN]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
|
06:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--02:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
05:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_partkey = p.p_partkey
|
|--01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB compact
| predicates: p.p_size = 15, p.p_type LIKE '%BRASS'
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q2_minimum_cost_supplier_tmp1, OVERWRITE=true]
| partitions=1
|
08:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: n.n_regionkey = r.r_regionkey
|
|--12:EXCHANGE [BROADCAST]
| |
| 04:SCAN HDFS [tpch.region r]
| partitions=1/1 size=384B
| predicates: r.r_name = 'EUROPE'
|
07:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--11:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--10:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
05:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_partkey = p.p_partkey
|
|--09:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB
| predicates: p.p_size = 15, p.p_type LIKE '%BRASS'
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
====
insert overwrite table tpch.q2_minimum_cost_supplier_tmp2
select
p_partkey,
min(ps_supplycost)
from tpch.q2_minimum_cost_supplier_tmp1
group by p_partkey
---- PLAN
WRITE TO HDFS [tpch.q2_minimum_cost_supplier_tmp2, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: min(ps_supplycost)
| group by: p_partkey
|
00:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp1]
partitions=1/1 size=107.06KB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q2_minimum_cost_supplier_tmp2, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: min(min(ps_supplycost))
| group by: p_partkey
|
02:EXCHANGE [HASH(p_partkey)]
|
01:AGGREGATE
| output: min(ps_supplycost)
| group by: p_partkey
|
00:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp1]
partitions=1/1 size=107.06KB
====
# Modifications: Fully qualified name of p_partkey column in the ORDER BY clause
select
t1.s_acctbal,
t1.s_name,
t1.n_name,
t1.p_partkey,
t1.p_mfgr,
t1.s_address,
t1.s_phone,
t1.s_comment
from tpch.q2_minimum_cost_supplier_tmp1 t1
join tpch.q2_minimum_cost_supplier_tmp2 t2
on (t1.p_partkey = t2.p_partkey and t1.ps_supplycost = t2.ps_min_supplycost)
order by
s_acctbal desc,
n_name,
s_name,
t1.p_partkey
limit 100
---- PLAN
03:TOP-N [LIMIT=100]
| order by: s_acctbal DESC, n_name ASC, s_name ASC, p_partkey ASC
|
02:HASH JOIN [INNER JOIN]
| hash predicates: t1.p_partkey = t2.p_partkey, t1.ps_supplycost = t2.ps_min_supplycost
|
|--01:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp2 t2]
| partitions=1/1 size=5.98KB compact
|
00:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp1 t1]
partitions=1/1 size=107.06KB
---- DISTRIBUTEDPLAN
05:MERGING-EXCHANGE [UNPARTITIONED]
| order by: s_acctbal DESC, n_name ASC, s_name ASC, p_partkey ASC
| limit: 100
|
03:TOP-N [LIMIT=100]
| order by: s_acctbal DESC, n_name ASC, s_name ASC, p_partkey ASC
|
02:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: t1.p_partkey = t2.p_partkey, t1.ps_supplycost = t2.ps_min_supplycost
|
|--04:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp2 t2]
| partitions=1/1 size=5.98KB
|
00:SCAN HDFS [tpch.q2_minimum_cost_supplier_tmp1 t1]
partitions=1/1 size=107.06KB
====
# Q3 - Shipping Priority Query
# Modifications: Added round() calls
select
l_orderkey,
round(sum(l_extendedprice * (1 - l_discount)), 5) as revenue,
o_orderdate,
o_shippriority
from tpch.lineitem l
join tpch.orders o
on (l.l_orderkey = o.o_orderkey)
join tpch.customer c
on (c.c_mktsegment = 'BUILDING' and c.c_custkey = o.o_custkey)
where
o_orderdate < '1995-03-15' and
l_shipdate > '1995-03-15'
group by
l_orderkey,
o_orderdate,
o_shippriority
order by
revenue desc,
o_orderdate
limit 10
---- PLAN
06:TOP-N [LIMIT=10]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC, o_orderdate ASC
|
05:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * (1 - l_discount))
| group by: l_orderkey, o_orderdate, o_shippriority
|
04:HASH JOIN [INNER JOIN]
| hash predicates: o.o_custkey = c.c_custkey
|
|--02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB compact
| predicates: c.c_mktsegment = 'BUILDING'
|
03:HASH JOIN [INNER JOIN]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o_orderdate < '1995-03-15'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l_shipdate > '1995-03-15'
---- DISTRIBUTEDPLAN
11:MERGING-EXCHANGE [UNPARTITIONED]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC, o_orderdate ASC
| limit: 10
|
06:TOP-N [LIMIT=10]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC, o_orderdate ASC
|
10:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * (1 - l_discount)))
| group by: l_orderkey, o_orderdate, o_shippriority
|
09:EXCHANGE [HASH(l_orderkey,o_orderdate,o_shippriority)]
|
05:AGGREGATE
| output: sum(l_extendedprice * (1 - l_discount))
| group by: l_orderkey, o_orderdate, o_shippriority
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o.o_custkey = c.c_custkey
|
|--08:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB
| predicates: c.c_mktsegment = 'BUILDING'
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--07:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o_orderdate < '1995-03-15'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l_shipdate > '1995-03-15'
====
# Q4 - Order Priority Checking Query
# Modifications: Converted selects from multiple tables to joins,
# rewrote 'exists' as an inner join + count(distinct), added
# limit
select
o_orderpriority,
count(distinct l_orderkey) as order_count
from tpch.lineitem l
inner join tpch.orders o
on (o.o_orderkey = l.l_orderkey and
l.l_commitdate < l.l_receiptdate)
where
o_orderdate >= '1993-07-01' and
o_orderdate < '1993-10-01'
group by
o_orderpriority
order by
o_orderpriority
limit 10
---- PLAN
05:TOP-N [LIMIT=10]
| order by: o_orderpriority ASC
|
04:AGGREGATE [MERGE FINALIZE]
| output: count(l_orderkey)
| group by: o_orderpriority
|
03:AGGREGATE
| group by: o_orderpriority, l_orderkey
|
02:HASH JOIN [INNER JOIN]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o_orderdate >= '1993-07-01', o_orderdate < '1993-10-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l.l_commitdate < l.l_receiptdate
---- DISTRIBUTEDPLAN
09:MERGING-EXCHANGE [UNPARTITIONED]
| order by: o_orderpriority ASC
| limit: 10
|
05:TOP-N [LIMIT=10]
| order by: o_orderpriority ASC
|
04:AGGREGATE [MERGE FINALIZE]
| output: count(l_orderkey)
| group by: o_orderpriority
|
08:AGGREGATE [MERGE]
| group by: o_orderpriority, l_orderkey
|
07:EXCHANGE [HASH(o_orderpriority)]
|
03:AGGREGATE
| group by: o_orderpriority, l_orderkey
|
02:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--06:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o_orderdate >= '1993-07-01', o_orderdate < '1993-10-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l.l_commitdate < l.l_receiptdate
====
# Q5 - Local Supplier Volume Query
# Modifications: Added round() call, converted selects from multiple tables
# to joins, added limit
select
n_name,
round(sum(l_extendedprice * (1 - l_discount)), 5) as revenue
from tpch.lineitem l
join tpch.orders o
on (l_orderkey = o_orderkey)
join tpch.supplier s
on (l_suppkey = s_suppkey)
join tpch.customer
on (c_nationkey = s_nationkey and c_custkey = o_custkey)
join tpch.nation
on (s_nationkey = n_nationkey)
join tpch.region
on (n_regionkey = r_regionkey)
where
r_name = 'ASIA'
and o_orderdate >= '1994-01-01'
and o_orderdate < '1995-01-01'
group by
n_name
order by
revenue desc
limit 100
---- PLAN
12:TOP-N [LIMIT=100]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
|
11:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * (1 - l_discount))
| group by: n_name
|
10:HASH JOIN [INNER JOIN]
| hash predicates: n_regionkey = r_regionkey
|
|--05:SCAN HDFS [tpch.region]
| partitions=1/1 size=384B compact
| predicates: r_name = 'ASIA'
|
09:HASH JOIN [INNER JOIN]
| hash predicates: s_nationkey = n_nationkey
|
|--04:SCAN HDFS [tpch.nation]
| partitions=1/1 size=2.15KB compact
|
08:HASH JOIN [INNER JOIN]
| hash predicates: s_nationkey = c_nationkey, o_custkey = c_custkey
|
|--03:SCAN HDFS [tpch.customer]
| partitions=1/1 size=23.08MB compact
|
07:HASH JOIN [INNER JOIN]
| hash predicates: l_suppkey = s_suppkey
|
|--02:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
06:HASH JOIN [INNER JOIN]
| hash predicates: l_orderkey = o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o_orderdate >= '1994-01-01', o_orderdate < '1995-01-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
20:MERGING-EXCHANGE [UNPARTITIONED]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
| limit: 100
|
12:TOP-N [LIMIT=100]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
|
19:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * (1 - l_discount)))
| group by: n_name
|
18:EXCHANGE [HASH(n_name)]
|
11:AGGREGATE
| output: sum(l_extendedprice * (1 - l_discount))
| group by: n_name
|
10:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: n_regionkey = r_regionkey
|
|--17:EXCHANGE [BROADCAST]
| |
| 05:SCAN HDFS [tpch.region]
| partitions=1/1 size=384B
| predicates: r_name = 'ASIA'
|
09:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s_nationkey = n_nationkey
|
|--16:EXCHANGE [BROADCAST]
| |
| 04:SCAN HDFS [tpch.nation]
| partitions=1/1 size=2.15KB
|
08:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s_nationkey = c_nationkey, o_custkey = c_custkey
|
|--15:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.customer]
| partitions=1/1 size=23.08MB
|
07:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l_suppkey = s_suppkey
|
|--14:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l_orderkey = o_orderkey
|
|--13:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o_orderdate >= '1994-01-01', o_orderdate < '1995-01-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
====
# Q6 - Forecasting Revenue Change Query
# Modifications: Added round() call
select round(sum(l_extendedprice * l_discount), 5) as revenue
from tpch.lineitem
where l_shipdate >= '1994-01-01' and
l_shipdate < '1995-01-01' and
l_discount >= 0.05 and
l_discount <= 0.07 and
l_quantity < 24
---- PLAN
01:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * l_discount)
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1994-01-01', l_shipdate < '1995-01-01', l_discount >= 0.05, l_discount <= 0.07, l_quantity < 24
---- DISTRIBUTEDPLAN
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * l_discount))
|
02:EXCHANGE [UNPARTITIONED]
|
01:AGGREGATE
| output: sum(l_extendedprice * l_discount)
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1994-01-01', l_shipdate < '1995-01-01', l_discount >= 0.05, l_discount <= 0.07, l_quantity < 24
====
# Q8 - National Market Share Query
# Modifications: Got rid of subquery, converted select from multiple tables to joins,
# added round() call
select
year(o_orderdate) as o_year,
round(sum(case when n2.n_name = 'BRAZIL' then l_extendedprice * (1 - l_discount)
else 0 end) / sum(l_extendedprice * (1 - l_discount)), 5) as mkt_share
from tpch.lineitem l
join tpch.orders o
on (l_orderkey = o_orderkey)
join tpch.part p
on (p_partkey = l_partkey)
join tpch.supplier s
on (s_suppkey = l_suppkey)
join tpch.customer c
on (o_custkey = c_custkey)
join tpch.nation n1
on (c_nationkey = n1.n_nationkey)
join tpch.region r
on (n1.n_regionkey = r_regionkey)
join tpch.nation n2
on (s_nationkey = n2.n_nationkey)
where
r_name = 'AMERICA' and
o_orderdate >= '1995-01-01' and
o_orderdate < '1996-12-31' and
p_type = 'ECONOMY ANODIZED STEEL'
group by
o_year
order by
o_year
limit 100
---- PLAN
16:TOP-N [LIMIT=100]
| order by: year(o_orderdate) ASC
|
15:AGGREGATE [FINALIZE]
| output: sum(CASE WHEN n2.n_name = 'BRAZIL' THEN l_extendedprice * (1 - l_discount) ELSE 0 END), sum(l_extendedprice * (1 - l_discount))
| group by: year(o_orderdate)
|
14:HASH JOIN [INNER JOIN]
| hash predicates: s_nationkey = n2.n_nationkey
|
|--07:SCAN HDFS [tpch.nation n2]
| partitions=1/1 size=2.15KB compact
|
13:HASH JOIN [INNER JOIN]
| hash predicates: n1.n_regionkey = r_regionkey
|
|--06:SCAN HDFS [tpch.region r]
| partitions=1/1 size=384B compact
| predicates: r_name = 'AMERICA'
|
12:HASH JOIN [INNER JOIN]
| hash predicates: c_nationkey = n1.n_nationkey
|
|--05:SCAN HDFS [tpch.nation n1]
| partitions=1/1 size=2.15KB compact
|
11:HASH JOIN [INNER JOIN]
| hash predicates: o_custkey = c_custkey
|
|--04:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB compact
|
10:HASH JOIN [INNER JOIN]
| hash predicates: l_suppkey = s_suppkey
|
|--03:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
09:HASH JOIN [INNER JOIN]
| hash predicates: l_partkey = p_partkey
|
|--02:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB compact
| predicates: p_type = 'ECONOMY ANODIZED STEEL'
|
08:HASH JOIN [INNER JOIN]
| hash predicates: l_orderkey = o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o_orderdate >= '1995-01-01', o_orderdate < '1996-12-31'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
26:MERGING-EXCHANGE [UNPARTITIONED]
| order by: year(o_orderdate) ASC
| limit: 100
|
16:TOP-N [LIMIT=100]
| order by: year(o_orderdate) ASC
|
25:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(CASE WHEN n2.n_name = 'BRAZIL' THEN l_extendedprice * (1 - l_discount) ELSE 0 END)), sum(sum(l_extendedprice * (1 - l_discount)))
| group by: year(o_orderdate)
|
24:EXCHANGE [HASH(year(o_orderdate))]
|
15:AGGREGATE
| output: sum(CASE WHEN n2.n_name = 'BRAZIL' THEN l_extendedprice * (1 - l_discount) ELSE 0 END), sum(l_extendedprice * (1 - l_discount))
| group by: year(o_orderdate)
|
14:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s_nationkey = n2.n_nationkey
|
|--23:EXCHANGE [BROADCAST]
| |
| 07:SCAN HDFS [tpch.nation n2]
| partitions=1/1 size=2.15KB
|
13:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: n1.n_regionkey = r_regionkey
|
|--22:EXCHANGE [BROADCAST]
| |
| 06:SCAN HDFS [tpch.region r]
| partitions=1/1 size=384B
| predicates: r_name = 'AMERICA'
|
12:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: c_nationkey = n1.n_nationkey
|
|--21:EXCHANGE [BROADCAST]
| |
| 05:SCAN HDFS [tpch.nation n1]
| partitions=1/1 size=2.15KB
|
11:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o_custkey = c_custkey
|
|--20:EXCHANGE [BROADCAST]
| |
| 04:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB
|
10:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l_suppkey = s_suppkey
|
|--19:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
09:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l_partkey = p_partkey
|
|--18:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB
| predicates: p_type = 'ECONOMY ANODIZED STEEL'
|
08:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l_orderkey = o_orderkey
|
|--17:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o_orderdate >= '1995-01-01', o_orderdate < '1996-12-31'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
====
# Q9 - Product Type Measure Query
# Modifications: Removed subquery, converted selects from multiple tables to joins,
# added limit, added round()
select
n.n_name as nation,
year(o.o_orderdate) as o_year,
round(sum(l.l_extendedprice * (1 - l.l_discount) -
ps.ps_supplycost * l.l_quantity), 2) as sum_profit
from tpch.lineitem l
join tpch.part p
on (p.p_partkey = l.l_partkey)
join tpch.orders o
on (o.o_orderkey = l.l_orderkey)
join tpch.partsupp ps
on (ps.ps_suppkey = l.l_suppkey and ps.ps_partkey = l.l_partkey)
join tpch.supplier s
on (s.s_suppkey = l.l_suppkey)
join tpch.nation n
on (s.s_nationkey = n.n_nationkey)
where
p.p_name like '%green%'
group by
n.n_name,
year(o.o_orderdate)
order by
nation,
o_year desc
limit 200
---- PLAN
12:TOP-N [LIMIT=200]
| order by: n.n_name ASC, year(o.o_orderdate) DESC
|
11:AGGREGATE [FINALIZE]
| output: sum(l.l_extendedprice * (1 - l.l_discount) - ps.ps_supplycost * l.l_quantity)
| group by: n.n_name, year(o.o_orderdate)
|
10:HASH JOIN [INNER JOIN]
| hash predicates: l.l_suppkey = ps.ps_suppkey, l.l_partkey = ps.ps_partkey
|
|--03:SCAN HDFS [tpch.partsupp ps]
| partitions=1/1 size=112.71MB compact
|
09:HASH JOIN [INNER JOIN]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--05:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
|
08:HASH JOIN [INNER JOIN]
| hash predicates: l.l_suppkey = s.s_suppkey
|
|--04:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
07:HASH JOIN [INNER JOIN]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
|
06:HASH JOIN [INNER JOIN]
| hash predicates: l.l_partkey = p.p_partkey
|
|--01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB compact
| predicates: p.p_name LIKE '%green%'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
20:MERGING-EXCHANGE [UNPARTITIONED]
| order by: n.n_name ASC, year(o.o_orderdate) DESC
| limit: 200
|
12:TOP-N [LIMIT=200]
| order by: n.n_name ASC, year(o.o_orderdate) DESC
|
19:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l.l_extendedprice * (1 - l.l_discount) - ps.ps_supplycost * l.l_quantity))
| group by: n.n_name, year(o.o_orderdate)
|
18:EXCHANGE [HASH(n.n_name,year(o.o_orderdate))]
|
11:AGGREGATE
| output: sum(l.l_extendedprice * (1 - l.l_discount) - ps.ps_supplycost * l.l_quantity)
| group by: n.n_name, year(o.o_orderdate)
|
10:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_suppkey = ps.ps_suppkey, l.l_partkey = ps.ps_partkey
|
|--17:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.partsupp ps]
| partitions=1/1 size=112.71MB
|
09:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--16:EXCHANGE [BROADCAST]
| |
| 05:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
|
08:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_suppkey = s.s_suppkey
|
|--15:EXCHANGE [BROADCAST]
| |
| 04:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
07:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--14:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_partkey = p.p_partkey
|
|--13:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB
| predicates: p.p_name LIKE '%green%'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
====
# Q10 - Returned Item Reporting Query
# Modifications: Cast c_acctbal to bigint due to float/double values not allowed in
# GROUP BY clause, added round() calls, converted select from multiple tables to
# joins
select
c_custkey,
c_name,
round(sum(l_extendedprice * (1 - l_discount)), 5) as revenue,
cast(c_acctbal as bigint) cast_c_acctbal,
n_name,
c_address,
c_phone,
c_comment
from tpch.lineitem l
join tpch.orders o
on (l.l_orderkey = o.o_orderkey)
join tpch.customer c
on (c.c_custkey = o.o_custkey)
join tpch.nation n
on (c.c_nationkey = n.n_nationkey)
where
o.o_orderdate >= '1993-10-01' and
o.o_orderdate < '1994-01-01' and
l.l_returnflag = 'R'
group by
c_custkey,
c_name,
cast(c_acctbal as bigint),
c_phone,
n_name,
c_address,
c_comment
order by
revenue desc
limit 20
---- PLAN
08:TOP-N [LIMIT=20]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
|
07:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * (1 - l_discount))
| group by: c_custkey, c_name, CAST(c_acctbal AS BIGINT), c_phone, n_name, c_address, c_comment
|
06:HASH JOIN [INNER JOIN]
| hash predicates: c.c_nationkey = n.n_nationkey
|
|--03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
|
05:HASH JOIN [INNER JOIN]
| hash predicates: o.o_custkey = c.c_custkey
|
|--02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB compact
|
04:HASH JOIN [INNER JOIN]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o.o_orderdate >= '1993-10-01', o.o_orderdate < '1994-01-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l.l_returnflag = 'R'
---- DISTRIBUTEDPLAN
14:MERGING-EXCHANGE [UNPARTITIONED]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
| limit: 20
|
08:TOP-N [LIMIT=20]
| order by: round(sum(l_extendedprice * (1 - l_discount)), 5) DESC
|
13:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * (1 - l_discount)))
| group by: c_custkey, c_name, CAST(c_acctbal AS BIGINT), c_phone, n_name, c_address, c_comment
|
12:EXCHANGE [HASH(c_custkey,c_name,CAST(c_acctbal AS BIGINT),c_phone,n_name,c_address,c_comment)]
|
07:AGGREGATE
| output: sum(l_extendedprice * (1 - l_discount))
| group by: c_custkey, c_name, CAST(c_acctbal AS BIGINT), c_phone, n_name, c_address, c_comment
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: c.c_nationkey = n.n_nationkey
|
|--11:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
|
05:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o.o_custkey = c.c_custkey
|
|--10:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--09:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o.o_orderdate >= '1993-10-01', o.o_orderdate < '1994-01-01'
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
predicates: l.l_returnflag = 'R'
====
# Q11 - Important Stock Identification
# TODO: Need to add part 2 of this query. It required some more advanced modifications.
insert overwrite table tpch.q11_part_tmp
select ps_partkey, sum(ps_supplycost * ps_availqty) as part_value
from tpch.nation n
join tpch.supplier s
on s.s_nationkey = n.n_nationkey and n.n_name = 'GERMANY'
join tpch.partsupp ps
on ps.ps_suppkey = s.s_suppkey
group by ps_partkey
---- PLAN
WRITE TO HDFS [tpch.q11_part_tmp, OVERWRITE=true]
| partitions=1
|
05:AGGREGATE [FINALIZE]
| output: sum(ps_supplycost * ps_availqty)
| group by: ps_partkey
|
04:HASH JOIN [INNER JOIN]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--00:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
| predicates: n.n_name = 'GERMANY'
|
03:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--01:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
02:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q11_part_tmp, OVERWRITE=true]
| partitions=1
|
09:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(ps_supplycost * ps_availqty))
| group by: ps_partkey
|
08:EXCHANGE [HASH(ps_partkey)]
|
05:AGGREGATE
| output: sum(ps_supplycost * ps_availqty)
| group by: ps_partkey
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--07:EXCHANGE [BROADCAST]
| |
| 00:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
| predicates: n.n_name = 'GERMANY'
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--06:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
02:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
====
insert overwrite table tpch.q11_sum_tmp
select sum(part_value) as total_value
from tpch.q11_part_tmp
---- PLAN
WRITE TO HDFS [tpch.q11_sum_tmp, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: sum(part_value)
|
00:SCAN HDFS [tpch.q11_part_tmp]
partitions=1/1 size=497.31KB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q11_sum_tmp, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(part_value))
|
02:EXCHANGE [UNPARTITIONED]
|
01:AGGREGATE
| output: sum(part_value)
|
00:SCAN HDFS [tpch.q11_part_tmp]
partitions=1/1 size=497.31KB
====
# Q12 - Shipping Mode and Order Priority Query
# Modifications: Converted select from multiple tables to joins, added limit
select l_shipmode,
sum(case
when o_orderpriority ='1-URGENT' or
o_orderpriority ='2-HIGH'
then 1
else 0
end
) as high_line_count,
sum(case
when o_orderpriority <> '1-URGENT' and
o_orderpriority <> '2-HIGH'
then 1
else 0
end
) as low_line_count
from tpch.lineitem l
join tpch.orders o
on (o.o_orderkey = l.l_orderkey and
l.l_commitdate < l.l_receiptdate and
l.l_shipdate < l.l_commitdate)
where
(l.l_shipmode = 'MAIL' or l.l_shipmode = 'SHIP') and
l.l_receiptdate >= '1994-01-01' and
l.l_receiptdate < '1995-01-01'
group by
l_shipmode
order by
l_shipmode
limit 10
---- PLAN
04:TOP-N [LIMIT=10]
| order by: l_shipmode ASC
|
03:AGGREGATE [FINALIZE]
| output: sum(CASE WHEN o_orderpriority = '1-URGENT' OR o_orderpriority = '2-HIGH' THEN 1 ELSE 0 END), sum(CASE WHEN o_orderpriority != '1-URGENT' AND o_orderpriority != '2-HIGH' THEN 1 ELSE 0 END)
| group by: l_shipmode
|
02:HASH JOIN [INNER JOIN]
| hash predicates: o.o_orderkey = l.l_orderkey
|
|--00:SCAN HDFS [tpch.lineitem l]
| partitions=1/1 size=718.94MB compact
| predicates: l.l_commitdate < l.l_receiptdate, l.l_shipdate < l.l_commitdate, (l.l_shipmode = 'MAIL' OR l.l_shipmode = 'SHIP'), l.l_receiptdate >= '1994-01-01', l.l_receiptdate < '1995-01-01'
|
01:SCAN HDFS [tpch.orders o]
partitions=1/1 size=162.56MB
---- DISTRIBUTEDPLAN
08:MERGING-EXCHANGE [UNPARTITIONED]
| order by: l_shipmode ASC
| limit: 10
|
04:TOP-N [LIMIT=10]
| order by: l_shipmode ASC
|
07:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(CASE WHEN o_orderpriority = '1-URGENT' OR o_orderpriority = '2-HIGH' THEN 1 ELSE 0 END)), sum(sum(CASE WHEN o_orderpriority != '1-URGENT' AND o_orderpriority != '2-HIGH' THEN 1 ELSE 0 END))
| group by: l_shipmode
|
06:EXCHANGE [HASH(l_shipmode)]
|
03:AGGREGATE
| output: sum(CASE WHEN o_orderpriority = '1-URGENT' OR o_orderpriority = '2-HIGH' THEN 1 ELSE 0 END), sum(CASE WHEN o_orderpriority != '1-URGENT' AND o_orderpriority != '2-HIGH' THEN 1 ELSE 0 END)
| group by: l_shipmode
|
02:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o.o_orderkey = l.l_orderkey
|
|--05:EXCHANGE [BROADCAST]
| |
| 00:SCAN HDFS [tpch.lineitem l]
| partitions=1/1 size=718.94MB
| predicates: l.l_commitdate < l.l_receiptdate, l.l_shipdate < l.l_commitdate, (l.l_shipmode = 'MAIL' OR l.l_shipmode = 'SHIP'), l.l_receiptdate >= '1994-01-01', l.l_receiptdate < '1995-01-01'
|
01:SCAN HDFS [tpch.orders o]
partitions=1/1 size=162.56MB
====
# Q13 - Customer Distribution Query
select
c_count,
count(1) as custdist
from
( select
c_custkey,
count(o_orderkey) as c_count
from tpch.orders o
right outer join tpch.customer c
on (c.c_custkey = o.o_custkey and o.o_comment not like '%special%requests%')
group by
c_custkey
) c_orders
group by
c_count
order by
custdist desc,
c_count desc
limit 100
---- PLAN
05:TOP-N [LIMIT=100]
| order by: count(1) DESC, c_count DESC
|
04:AGGREGATE [FINALIZE]
| output: count(1)
| group by: count(o_orderkey)
|
03:AGGREGATE [FINALIZE]
| output: count(o_orderkey)
| group by: c_custkey
|
02:HASH JOIN [RIGHT OUTER JOIN]
| hash predicates: o.o_custkey = c.c_custkey
|
|--01:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB compact
|
00:SCAN HDFS [tpch.orders o]
partitions=1/1 size=162.56MB
predicates: NOT o.o_comment LIKE '%special%requests%'
---- DISTRIBUTEDPLAN
12:MERGING-EXCHANGE [UNPARTITIONED]
| order by: count(1) DESC, c_count DESC
| limit: 100
|
05:TOP-N [LIMIT=100]
| order by: count(1) DESC, c_count DESC
|
11:AGGREGATE [MERGE FINALIZE]
| output: sum(count(1))
| group by: c_count
|
10:EXCHANGE [HASH(c_count)]
|
04:AGGREGATE
| output: count(1)
| group by: count(o_orderkey)
|
09:AGGREGATE [MERGE FINALIZE]
| output: sum(count(o_orderkey))
| group by: c_custkey
|
08:EXCHANGE [HASH(c_custkey)]
|
03:AGGREGATE
| output: count(o_orderkey)
| group by: c_custkey
|
02:HASH JOIN [RIGHT OUTER JOIN, PARTITIONED]
| hash predicates: o.o_custkey = c.c_custkey
|
|--07:EXCHANGE [HASH(c.c_custkey)]
| |
| 01:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB
|
06:EXCHANGE [HASH(o.o_custkey)]
|
00:SCAN HDFS [tpch.orders o]
partitions=1/1 size=162.56MB
predicates: NOT o.o_comment LIKE '%special%requests%'
====
# Q14 - Promotion Effect
select
round(100.00 * sum(case when p_type like 'PROMO%' then l_extendedprice*(1-l_discount)
else 0.0
end
) / sum(l_extendedprice * (1 - l_discount)), 5) as promo_revenue
from tpch.lineitem l
join tpch.part p
on l.l_partkey = p.p_partkey and
l.l_shipdate >= '1995-09-01' and
l.l_shipdate < '1995-10-01'
---- PLAN
03:AGGREGATE [FINALIZE]
| output: sum(CASE WHEN p_type LIKE 'PROMO%' THEN l_extendedprice * (1 - l_discount) ELSE 0.0 END), sum(l_extendedprice * (1 - l_discount))
|
02:HASH JOIN [INNER JOIN]
| hash predicates: p.p_partkey = l.l_partkey
|
|--00:SCAN HDFS [tpch.lineitem l]
| partitions=1/1 size=718.94MB compact
| predicates: l.l_shipdate >= '1995-09-01', l.l_shipdate < '1995-10-01'
|
01:SCAN HDFS [tpch.part p]
partitions=1/1 size=22.83MB
---- DISTRIBUTEDPLAN
06:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(CASE WHEN p_type LIKE 'PROMO%' THEN l_extendedprice * (1 - l_discount) ELSE 0.0 END)), sum(sum(l_extendedprice * (1 - l_discount)))
|
05:EXCHANGE [UNPARTITIONED]
|
03:AGGREGATE
| output: sum(CASE WHEN p_type LIKE 'PROMO%' THEN l_extendedprice * (1 - l_discount) ELSE 0.0 END), sum(l_extendedprice * (1 - l_discount))
|
02:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: p.p_partkey = l.l_partkey
|
|--04:EXCHANGE [BROADCAST]
| |
| 00:SCAN HDFS [tpch.lineitem l]
| partitions=1/1 size=718.94MB
| predicates: l.l_shipdate >= '1995-09-01', l.l_shipdate < '1995-10-01'
|
01:SCAN HDFS [tpch.part p]
partitions=1/1 size=22.83MB
====
# Q15 - Top Supplier Query
insert overwrite table tpch.revenue
select
l_suppkey as supplier_no,
sum(l_extendedprice * (1 - l_discount)) as total_revenue
from tpch.lineitem
where l_shipdate >= '1996-01-01' and l_shipdate < '1996-04-01'
group by l_suppkey
---- PLAN
WRITE TO HDFS [tpch.revenue, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * (1 - l_discount))
| group by: l_suppkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1996-01-01', l_shipdate < '1996-04-01'
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.revenue, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * (1 - l_discount)))
| group by: l_suppkey
|
02:EXCHANGE [HASH(l_suppkey)]
|
01:AGGREGATE
| output: sum(l_extendedprice * (1 - l_discount))
| group by: l_suppkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1996-01-01', l_shipdate < '1996-04-01'
====
insert overwrite table tpch.max_revenue
select max(total_revenue)
from tpch.revenue
---- PLAN
WRITE TO HDFS [tpch.max_revenue, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: max(total_revenue)
|
00:SCAN HDFS [tpch.revenue]
partitions=1/1 size=166.89KB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.max_revenue, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: max(max(total_revenue))
|
02:EXCHANGE [UNPARTITIONED]
|
01:AGGREGATE
| output: max(total_revenue)
|
00:SCAN HDFS [tpch.revenue]
partitions=1/1 size=166.89KB
====
# Modifications - Added limit
select
s_suppkey,
s_name,
s_address,
s_phone,
total_revenue
from tpch.supplier s
join tpch.revenue r
on (s.s_suppkey = r.supplier_no)
join tpch.max_revenue m
on (r.total_revenue = m.max_revenue)
order by s_suppkey
limit 100
---- PLAN
05:TOP-N [LIMIT=100]
| order by: s_suppkey ASC
|
04:HASH JOIN [INNER JOIN]
| hash predicates: r.total_revenue = m.max_revenue
|
|--02:SCAN HDFS [tpch.max_revenue m]
| partitions=1/1 size=13B compact
|
03:HASH JOIN [INNER JOIN]
| hash predicates: s.s_suppkey = r.supplier_no
|
|--01:SCAN HDFS [tpch.revenue r]
| partitions=1/1 size=166.89KB compact
|
00:SCAN HDFS [tpch.supplier s]
partitions=1/1 size=1.33MB
---- DISTRIBUTEDPLAN
08:MERGING-EXCHANGE [UNPARTITIONED]
| order by: s_suppkey ASC
| limit: 100
|
05:TOP-N [LIMIT=100]
| order by: s_suppkey ASC
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: r.total_revenue = m.max_revenue
|
|--07:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.max_revenue m]
| partitions=1/1 size=13B
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_suppkey = r.supplier_no
|
|--06:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.revenue r]
| partitions=1/1 size=166.89KB
|
00:SCAN HDFS [tpch.supplier s]
partitions=1/1 size=1.33MB
====
# Q16 - Parts/Supplier Relation Query
insert overwrite table tpch.supplier_tmp
select s_suppkey
from tpch.supplier
where not s_comment like '%Customer%Complaints%'
---- PLAN
WRITE TO HDFS [tpch.supplier_tmp, OVERWRITE=true]
| partitions=1
|
00:SCAN HDFS [tpch.supplier]
partitions=1/1 size=1.33MB
predicates: NOT s_comment LIKE '%Customer%Complaints%'
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.supplier_tmp, OVERWRITE=true]
| partitions=1
|
00:SCAN HDFS [tpch.supplier]
partitions=1/1 size=1.33MB
predicates: NOT s_comment LIKE '%Customer%Complaints%'
====
insert overwrite table tpch.q16_tmp
select p_brand, p_type, p_size, ps_suppkey
from tpch.partsupp ps
join tpch.part p
on p.p_partkey = ps.ps_partkey and
p.p_brand <> 'Brand#45' and
not p.p_type like 'MEDIUM POLISHED%'
join tpch.supplier_tmp s
on ps.ps_suppkey = s.s_suppkey
---- PLAN
WRITE TO HDFS [tpch.q16_tmp, OVERWRITE=true]
| partitions=1
|
04:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--02:SCAN HDFS [tpch.supplier_tmp s]
| partitions=1/1 size=47.73KB compact
|
03:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_partkey = p.p_partkey
|
|--01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB compact
| predicates: p.p_brand != 'Brand#45', NOT p.p_type LIKE 'MEDIUM POLISHED%'
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q16_tmp, OVERWRITE=true]
| partitions=1
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_suppkey = s.s_suppkey
|
|--06:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.supplier_tmp s]
| partitions=1/1 size=47.73KB
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_partkey = p.p_partkey
|
|--05:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB
| predicates: p.p_brand != 'Brand#45', NOT p.p_type LIKE 'MEDIUM POLISHED%'
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
====
# Modifications: Added limit
select p_brand, p_type, p_size, count(distinct ps_suppkey) as supplier_cnt
from
( select * from tpch.q16_tmp
where p_size = 49 or p_size = 14 or
p_size = 23 or p_size = 45 or
p_size = 19 or p_size = 3 or
p_size = 36 or p_size = 9
) q16_all
group by p_brand, p_type, p_size
order by supplier_cnt desc, p_brand, p_type, p_size
limit 1000
---- PLAN
03:TOP-N [LIMIT=1000]
| order by: count(ps_suppkey) DESC, p_brand ASC, p_type ASC, p_size ASC
|
02:AGGREGATE [MERGE FINALIZE]
| output: count(ps_suppkey)
| group by: p_brand, p_type, p_size
|
01:AGGREGATE
| group by: tpch.q16_tmp.p_brand, tpch.q16_tmp.p_type, tpch.q16_tmp.p_size, tpch.q16_tmp.ps_suppkey
|
00:SCAN HDFS [tpch.q16_tmp]
partitions=1/1 size=27.10MB
predicates: p_size = 49 OR p_size = 14 OR p_size = 23 OR p_size = 45 OR p_size = 19 OR p_size = 3 OR p_size = 36 OR p_size = 9
---- DISTRIBUTEDPLAN
06:MERGING-EXCHANGE [UNPARTITIONED]
| order by: count(ps_suppkey) DESC, p_brand ASC, p_type ASC, p_size ASC
| limit: 1000
|
03:TOP-N [LIMIT=1000]
| order by: count(ps_suppkey) DESC, p_brand ASC, p_type ASC, p_size ASC
|
02:AGGREGATE [MERGE FINALIZE]
| output: count(ps_suppkey)
| group by: p_brand, p_type, p_size
|
05:AGGREGATE [MERGE]
| group by: p_brand, p_type, p_size, ps_suppkey
|
04:EXCHANGE [HASH(p_brand,p_type,p_size)]
|
01:AGGREGATE
| group by: tpch.q16_tmp.p_brand, tpch.q16_tmp.p_type, tpch.q16_tmp.p_size, tpch.q16_tmp.ps_suppkey
|
00:SCAN HDFS [tpch.q16_tmp]
partitions=1/1 size=27.10MB
predicates: p_size = 49 OR p_size = 14 OR p_size = 23 OR p_size = 45 OR p_size = 19 OR p_size = 3 OR p_size = 36 OR p_size = 9
====
# Q17 - Small-Quantity-Order Revenue Query
insert overwrite table tpch.lineitem_tmp
select l_partkey as t_partkey, 0.2 * avg(l_quantity) as t_avg_quantity
from tpch.lineitem
group by l_partkey
====
# Modifications: Converted selects from multiple tables to joins,
# added round() call, removed subquery
select round(sum(l_extendedprice) / 7.0, 5) as avg_yearly
from tpch.lineitem l
join tpch.part p
on (p.p_partkey = l.l_partkey)
join tpch.lineitem_tmp lt
on (lt.t_partkey = p.p_partkey)
where
p.p_brand = 'Brand#23' and
p.p_container = 'MED BOX' and
l.l_quantity < lt.t_avg_quantity
---- PLAN
---- DISTRIBUTEDPLAN
====
# Q18 - Large Value Customer Query
insert overwrite table tpch.q18_tmp
select l_orderkey, sum(l_quantity) as t_sum_quantity
from tpch.lineitem
group by l_orderkey
---- PLAN
WRITE TO HDFS [tpch.q18_tmp, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: sum(l_quantity)
| group by: l_orderkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q18_tmp, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_quantity))
| group by: l_orderkey
|
02:EXCHANGE [HASH(l_orderkey)]
|
01:AGGREGATE
| output: sum(l_quantity)
| group by: l_orderkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
====
# Modifications: Cast o_totalprice column due to float/double values not
# allowed in GROUP BY.
select
c_name,
c_custkey,
o_orderkey,
o_orderdate,
cast(o_totalprice as bigint) as total_price_bigint,
round(sum(l_quantity), 5)
from tpch.lineitem l
join tpch.orders o
on (o.o_orderkey = l.l_orderkey)
join tpch.customer c
on (c.c_custkey = o.o_custkey)
join tpch.q18_tmp t
on (o.o_orderkey = t.l_orderkey and t.t_sum_quantity > 300)
group by
c_name,
c_custkey,
o_orderkey,
o_orderdate,
cast(o_totalprice as bigint)
order by
total_price_bigint,
o_orderdate
limit 100
---- PLAN
08:TOP-N [LIMIT=100]
| order by: CAST(o_totalprice AS BIGINT) ASC, o_orderdate ASC
|
07:AGGREGATE [FINALIZE]
| output: sum(l_quantity)
| group by: c_name, c_custkey, o_orderkey, o_orderdate, CAST(o_totalprice AS BIGINT)
|
06:HASH JOIN [INNER JOIN]
| hash predicates: o.o_orderkey = t.l_orderkey
|
|--03:SCAN HDFS [tpch.q18_tmp t]
| partitions=1/1 size=20.43MB compact
| predicates: t.t_sum_quantity > 300
|
05:HASH JOIN [INNER JOIN]
| hash predicates: o.o_custkey = c.c_custkey
|
|--02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB compact
|
04:HASH JOIN [INNER JOIN]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
15:MERGING-EXCHANGE [UNPARTITIONED]
| order by: CAST(o_totalprice AS BIGINT) ASC, o_orderdate ASC
| limit: 100
|
08:TOP-N [LIMIT=100]
| order by: CAST(o_totalprice AS BIGINT) ASC, o_orderdate ASC
|
14:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_quantity))
| group by: c_name, c_custkey, o_orderkey, o_orderdate, CAST(o_totalprice AS BIGINT)
|
13:EXCHANGE [HASH(c_name,c_custkey,o_orderkey,o_orderdate,CAST(o_totalprice AS BIGINT))]
|
07:AGGREGATE
| output: sum(l_quantity)
| group by: c_name, c_custkey, o_orderkey, o_orderdate, CAST(o_totalprice AS BIGINT)
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o.o_orderkey = t.l_orderkey
|
|--12:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.q18_tmp t]
| partitions=1/1 size=20.43MB
| predicates: t.t_sum_quantity > 300
|
05:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: o.o_custkey = c.c_custkey
|
|--11:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.customer c]
| partitions=1/1 size=23.08MB
|
04:HASH JOIN [INNER JOIN, PARTITIONED]
| hash predicates: l.l_orderkey = o.o_orderkey
|
|--10:EXCHANGE [HASH(o.o_orderkey)]
| |
| 01:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
|
09:EXCHANGE [HASH(l.l_orderkey)]
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
====
# Q19 - Discounted Revenue Query
# Modifications: Added round() calls
select round(sum(l_extendedprice * (1 - l_discount) ), 5) as revenue
from tpch.lineitem l
join tpch.part p
on p.p_partkey = l.l_partkey
where
(
p_brand = 'Brand#12'
and (p_container LIKE 'SM CASE' or
p_container LIKE 'SM BOX' or
p_container LIKE 'SM PACK' or
p_container LIKE 'SM PKG')
and l_quantity >= 1 and l_quantity <= 11
and p_size >= 1 and p_size <= 5
and (l_shipmode LIKE 'AIR' or
l_shipmode LIKE 'AIR REG')
and l_shipinstruct = 'DELIVER IN PERSON'
)
or
(
p_brand = 'Brand#23'
and (p_container LIKE 'MED BAG' or
p_container LIKE 'MED BOX' or
p_container LIKE 'MED PKG' or
p_container LIKE 'MED PACK')
and l_quantity >= 10 and l_quantity <= 20
and p_size >= 1 and p_size <= 10
and (l_shipmode LIKE 'AIR' or
l_shipmode LIKE 'AIR REG')
and l_shipinstruct = 'DELIVER IN PERSON'
)
or
(
p_brand = 'Brand#34'
and (p_container LIKE 'LG BAG' or
p_container LIKE 'LG BOX' or
p_container LIKE 'LG PKG' or
p_container LIKE 'LG PACK')
and l_quantity >= 20 and l_quantity <= 30
and p_size >= 1 and p_size <= 15
and (l_shipmode LIKE 'AIR' or
l_shipmode LIKE 'AIR REG')
and l_shipinstruct = 'DELIVER IN PERSON'
)
---- PLAN
03:AGGREGATE [FINALIZE]
| output: sum(l_extendedprice * (1 - l_discount))
|
02:HASH JOIN [INNER JOIN]
| hash predicates: l.l_partkey = p.p_partkey
| other predicates: (p_brand = 'Brand#12' AND (p_container LIKE 'SM CASE' OR p_container LIKE 'SM BOX' OR p_container LIKE 'SM PACK' OR p_container LIKE 'SM PKG') AND l_quantity >= 1 AND l_quantity <= 11 AND p_size >= 1 AND p_size <= 5 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON') OR (p_brand = 'Brand#23' AND (p_container LIKE 'MED BAG' OR p_container LIKE 'MED BOX' OR p_container LIKE 'MED PKG' OR p_container LIKE 'MED PACK') AND l_quantity >= 10 AND l_quantity <= 20 AND p_size >= 1 AND p_size <= 10 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON') OR (p_brand = 'Brand#34' AND (p_container LIKE 'LG BAG' OR p_container LIKE 'LG BOX' OR p_container LIKE 'LG PKG' OR p_container LIKE 'LG PACK') AND l_quantity >= 20 AND l_quantity <= 30 AND p_size >= 1 AND p_size <= 15 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON')
|
|--01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB compact
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
06:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_extendedprice * (1 - l_discount)))
|
05:EXCHANGE [UNPARTITIONED]
|
03:AGGREGATE
| output: sum(l_extendedprice * (1 - l_discount))
|
02:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l.l_partkey = p.p_partkey
| other predicates: (p_brand = 'Brand#12' AND (p_container LIKE 'SM CASE' OR p_container LIKE 'SM BOX' OR p_container LIKE 'SM PACK' OR p_container LIKE 'SM PKG') AND l_quantity >= 1 AND l_quantity <= 11 AND p_size >= 1 AND p_size <= 5 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON') OR (p_brand = 'Brand#23' AND (p_container LIKE 'MED BAG' OR p_container LIKE 'MED BOX' OR p_container LIKE 'MED PKG' OR p_container LIKE 'MED PACK') AND l_quantity >= 10 AND l_quantity <= 20 AND p_size >= 1 AND p_size <= 10 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON') OR (p_brand = 'Brand#34' AND (p_container LIKE 'LG BAG' OR p_container LIKE 'LG BOX' OR p_container LIKE 'LG PKG' OR p_container LIKE 'LG PACK') AND l_quantity >= 20 AND l_quantity <= 30 AND p_size >= 1 AND p_size <= 15 AND (l_shipmode LIKE 'AIR' OR l_shipmode LIKE 'AIR REG') AND l_shipinstruct = 'DELIVER IN PERSON')
|
|--04:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.part p]
| partitions=1/1 size=22.83MB
|
00:SCAN HDFS [tpch.lineitem l]
partitions=1/1 size=718.94MB
====
# QUERY_NAME : TPCH-Q20_QUERY_1
# Q20 - Potential Part Promotion Query
insert overwrite table tpch.q20_tmp1
select distinct p_partkey
from tpch.part
where p_name like 'forest%'
---- PLAN
WRITE TO HDFS [tpch.q20_tmp1, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| group by: p_partkey
|
00:SCAN HDFS [tpch.part]
partitions=1/1 size=22.83MB
predicates: p_name LIKE 'forest%'
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q20_tmp1, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| group by: p_partkey
|
02:EXCHANGE [HASH(p_partkey)]
|
01:AGGREGATE
| group by: p_partkey
|
00:SCAN HDFS [tpch.part]
partitions=1/1 size=22.83MB
predicates: p_name LIKE 'forest%'
====
# QUERY_NAME : TPCH-Q20_QUERY_2
insert overwrite table tpch.q20_tmp2
select
l_partkey,
l_suppkey,
0.5 * sum(l_quantity)
from tpch.lineitem
where
l_shipdate >= '1994-01-01' and
l_shipdate < '1995-01-01'
group by
l_partkey,
l_suppkey
---- PLAN
WRITE TO HDFS [tpch.q20_tmp2, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: sum(l_quantity)
| group by: l_partkey, l_suppkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1994-01-01', l_shipdate < '1995-01-01'
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q20_tmp2, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(l_quantity))
| group by: l_partkey, l_suppkey
|
02:EXCHANGE [HASH(l_partkey,l_suppkey)]
|
01:AGGREGATE
| output: sum(l_quantity)
| group by: l_partkey, l_suppkey
|
00:SCAN HDFS [tpch.lineitem]
partitions=1/1 size=718.94MB
predicates: l_shipdate >= '1994-01-01', l_shipdate < '1995-01-01'
====
# QUERY_NAME : TPCH-Q20_QUERY_3
insert overwrite table tpch.q20_tmp3
select
ps_suppkey,
ps_availqty,
sum_quantity
from tpch.partsupp ps
join tpch.q20_tmp2 t2
on (ps.ps_partkey = t2.l_partkey and ps.ps_suppkey = t2.l_suppkey)
join tpch.q20_tmp1 t1
on (ps.ps_partkey = t1.p_partkey)
---- PLAN
WRITE TO HDFS [tpch.q20_tmp3, OVERWRITE=true]
| partitions=1
|
04:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_partkey = t1.p_partkey
|
|--02:SCAN HDFS [tpch.q20_tmp1 t1]
| partitions=1/1 size=13.42KB compact
|
03:HASH JOIN [INNER JOIN]
| hash predicates: ps.ps_partkey = t2.l_partkey, ps.ps_suppkey = t2.l_suppkey
|
|--01:SCAN HDFS [tpch.q20_tmp2 t2]
| partitions=1/1 size=9.38MB compact
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q20_tmp3, OVERWRITE=true]
| partitions=1
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_partkey = t1.p_partkey
|
|--06:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.q20_tmp1 t1]
| partitions=1/1 size=13.42KB
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: ps.ps_partkey = t2.l_partkey, ps.ps_suppkey = t2.l_suppkey
|
|--05:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.q20_tmp2 t2]
| partitions=1/1 size=9.38MB
|
00:SCAN HDFS [tpch.partsupp ps]
partitions=1/1 size=112.71MB
====
# QUERY_NAME : TPCH-Q20_QUERY_4
# Modified to use subquery to work around IMP-127
insert overwrite table tpch.q20_tmp4
select ps_suppkey
from tpch.q20_tmp3
where ps_availqty > sum_quantity
group by ps_suppkey
---- PLAN
WRITE TO HDFS [tpch.q20_tmp4, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| group by: ps_suppkey
|
00:SCAN HDFS [tpch.q20_tmp3]
partitions=1/1 size=94.49KB
predicates: ps_availqty > sum_quantity
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q20_tmp4, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| group by: ps_suppkey
|
02:EXCHANGE [HASH(ps_suppkey)]
|
01:AGGREGATE
| group by: ps_suppkey
|
00:SCAN HDFS [tpch.q20_tmp3]
partitions=1/1 size=94.49KB
predicates: ps_availqty > sum_quantity
====
# QUERY_NAME : TPCH-Q20_QUERY_5
# Modifications: Added limit
select
s_name,
s_address
from tpch.supplier s
join tpch.nation n
on (s.s_nationkey = n.n_nationkey and
n.n_name = 'CANADA')
join tpch.q20_tmp4 t4
on (s.s_suppkey = t4.ps_suppkey)
order by
s_name
limit 100
---- PLAN
05:TOP-N [LIMIT=100]
| order by: s_name ASC
|
04:HASH JOIN [INNER JOIN]
| hash predicates: s.s_suppkey = t4.ps_suppkey
|
|--02:SCAN HDFS [tpch.q20_tmp4 t4]
| partitions=1/1 size=20.98KB compact
|
03:HASH JOIN [INNER JOIN]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--01:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
| predicates: n.n_name = 'CANADA'
|
00:SCAN HDFS [tpch.supplier s]
partitions=1/1 size=1.33MB
---- DISTRIBUTEDPLAN
08:MERGING-EXCHANGE [UNPARTITIONED]
| order by: s_name ASC
| limit: 100
|
05:TOP-N [LIMIT=100]
| order by: s_name ASC
|
04:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_suppkey = t4.ps_suppkey
|
|--07:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.q20_tmp4 t4]
| partitions=1/1 size=20.98KB
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--06:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
| predicates: n.n_name = 'CANADA'
|
00:SCAN HDFS [tpch.supplier s]
partitions=1/1 size=1.33MB
====
# QUERY_NAME : TPCH-Q21
# Q21 - Suppliers Who Kept Orders Waiting Query
# Modifications: Converted 'EXISTS' into LEFT SEMI JOIN, converted
# NOT EXISTS into LEFT OUTER JOIN, changed selects from multiple tables to
# joins, added limit
select
s_name,
count(*) as numwait
from tpch.lineitem l1
join tpch.supplier s
on (s.s_suppkey = l1.l_suppkey)
join tpch.orders o
on (o.o_orderkey = l1.l_orderkey)
join tpch.nation n
on (s.s_nationkey = n.n_nationkey)
left semi join tpch.lineitem l2
on (l2.l_orderkey = l1.l_orderkey and
l2.l_suppkey <> l1.l_suppkey)
left outer join tpch.lineitem l3
on (l3.l_orderkey = l1.l_orderkey and
l3.l_suppkey <> l1.l_suppkey and
l3.l_receiptdate > l3.l_commitdate)
where
l3.l_orderkey is null and
n_name = 'SAUDI ARABIA' and
o_orderstatus = 'F'
group by
s_name
order by
numwait desc,
s_name
limit 100
---- PLAN
12:TOP-N [LIMIT=100]
| order by: count(*) DESC, s_name ASC
|
11:AGGREGATE [FINALIZE]
| output: count(*)
| group by: s_name
|
10:HASH JOIN [LEFT OUTER JOIN]
| hash predicates: l1.l_orderkey = l3.l_orderkey
| other join predicates: l3.l_suppkey != l1.l_suppkey
| other predicates: l3.l_orderkey IS NULL
|
|--05:SCAN HDFS [tpch.lineitem l3]
| partitions=1/1 size=718.94MB compact
| predicates: l3.l_receiptdate > l3.l_commitdate
|
09:HASH JOIN [LEFT SEMI JOIN]
| hash predicates: l1.l_orderkey = l2.l_orderkey
| other predicates: l2.l_suppkey != l1.l_suppkey
|
|--04:SCAN HDFS [tpch.lineitem l2]
| partitions=1/1 size=718.94MB compact
|
08:HASH JOIN [INNER JOIN]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB compact
| predicates: n_name = 'SAUDI ARABIA'
|
07:HASH JOIN [INNER JOIN]
| hash predicates: l1.l_orderkey = o.o_orderkey
|
|--02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
| predicates: o_orderstatus = 'F'
|
06:HASH JOIN [INNER JOIN]
| hash predicates: l1.l_suppkey = s.s_suppkey
|
|--01:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB compact
|
00:SCAN HDFS [tpch.lineitem l1]
partitions=1/1 size=718.94MB
---- DISTRIBUTEDPLAN
20:MERGING-EXCHANGE [UNPARTITIONED]
| order by: count(*) DESC, s_name ASC
| limit: 100
|
12:TOP-N [LIMIT=100]
| order by: count(*) DESC, s_name ASC
|
19:AGGREGATE [MERGE FINALIZE]
| output: sum(count(*))
| group by: s_name
|
18:EXCHANGE [HASH(s_name)]
|
11:AGGREGATE
| output: count(*)
| group by: s_name
|
10:HASH JOIN [LEFT OUTER JOIN, BROADCAST]
| hash predicates: l1.l_orderkey = l3.l_orderkey
| other join predicates: l3.l_suppkey != l1.l_suppkey
| other predicates: l3.l_orderkey IS NULL
|
|--17:EXCHANGE [BROADCAST]
| |
| 05:SCAN HDFS [tpch.lineitem l3]
| partitions=1/1 size=718.94MB
| predicates: l3.l_receiptdate > l3.l_commitdate
|
09:HASH JOIN [LEFT SEMI JOIN, BROADCAST]
| hash predicates: l1.l_orderkey = l2.l_orderkey
| other predicates: l2.l_suppkey != l1.l_suppkey
|
|--16:EXCHANGE [BROADCAST]
| |
| 04:SCAN HDFS [tpch.lineitem l2]
| partitions=1/1 size=718.94MB
|
08:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: s.s_nationkey = n.n_nationkey
|
|--15:EXCHANGE [BROADCAST]
| |
| 03:SCAN HDFS [tpch.nation n]
| partitions=1/1 size=2.15KB
| predicates: n_name = 'SAUDI ARABIA'
|
07:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l1.l_orderkey = o.o_orderkey
|
|--14:EXCHANGE [BROADCAST]
| |
| 02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
| predicates: o_orderstatus = 'F'
|
06:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: l1.l_suppkey = s.s_suppkey
|
|--13:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.supplier s]
| partitions=1/1 size=1.33MB
|
00:SCAN HDFS [tpch.lineitem l1]
partitions=1/1 size=718.94MB
====
# QUERY_NAME : TPCH-Q22_QUERY_1
# Q22 - Global Sales Opportunity Query
# Modifications: Added a cust_name_char column that will always have
# a constant value ('C') so that we can do a join between this table
# in the main query. This was needed because we only support equi-joins
# and had to have a column to join on.
insert overwrite table tpch.q22_customer_tmp1
select
avg(c_acctbal) avg_acctbal,
substr(c_name, 1, 1) as cust_name_char
from tpch.customer c
where
c.c_acctbal > 0.00 and
(substr(c.c_phone, 1, 2) = '13' or
substr(c.c_phone, 1, 2) = '31' or
substr(c.c_phone, 1, 2) = '23' or
substr(c.c_phone, 1, 2) = '29' or
substr(c.c_phone, 1, 2) = '30' or
substr(c.c_phone, 1, 2) = '18' or
substr(c.c_phone, 1, 2) = '17')
group by
substr(c_name, 1, 1)
---- PLAN
WRITE TO HDFS [tpch.q22_customer_tmp1, OVERWRITE=true]
| partitions=1
|
01:AGGREGATE [FINALIZE]
| output: sum(c_acctbal), count(c_acctbal)
| group by: substr(c_name, 1, 1)
|
00:SCAN HDFS [tpch.customer c]
partitions=1/1 size=23.08MB
predicates: c.c_acctbal > 0.00, (substr(c.c_phone, 1, 2) = '13' OR substr(c.c_phone, 1, 2) = '31' OR substr(c.c_phone, 1, 2) = '23' OR substr(c.c_phone, 1, 2) = '29' OR substr(c.c_phone, 1, 2) = '30' OR substr(c.c_phone, 1, 2) = '18' OR substr(c.c_phone, 1, 2) = '17')
---- DISTRIBUTEDPLAN
WRITE TO HDFS [tpch.q22_customer_tmp1, OVERWRITE=true]
| partitions=1
|
03:AGGREGATE [MERGE FINALIZE]
| output: sum(sum(c_acctbal)), sum(count(c_acctbal))
| group by: substr(c_name, 1, 1)
|
02:EXCHANGE [HASH(substr(c_name, 1, 1))]
|
01:AGGREGATE
| output: sum(c_acctbal), count(c_acctbal)
| group by: substr(c_name, 1, 1)
|
00:SCAN HDFS [tpch.customer c]
partitions=1/1 size=23.08MB
predicates: c.c_acctbal > 0.00, (substr(c.c_phone, 1, 2) = '13' OR substr(c.c_phone, 1, 2) = '31' OR substr(c.c_phone, 1, 2) = '23' OR substr(c.c_phone, 1, 2) = '29' OR substr(c.c_phone, 1, 2) = '30' OR substr(c.c_phone, 1, 2) = '18' OR substr(c.c_phone, 1, 2) = '17')
====
# QUERY_NAME : TPCH-Q22_QUERY_2
# Modifications: Updated to use LEFT OUTER JOIN instead of NOT EXISTS,
# removed subquery by pushing aggregation up a level, added temp table
# to remove another subquery that computed the average account balance,
# added limit, added round()
select
substring(c_phone, 1, 2) as cntrycode,
count(*) as numcust,
round(sum(c_acctbal), 4) as totacctbal
from tpch.customer c
join tpch.q22_customer_tmp1 ct
on (substr(c.c_name, 1, 1) = ct.cust_name_char)
left outer join tpch.orders o
on (o.o_custkey = c.c_custkey)
where
o_custkey is null and
c.c_acctbal > ct.avg_acctbal and
(substr(c.c_phone, 1, 2) = '13' or
substr(c.c_phone, 1, 2) = '31' or
substr(c.c_phone, 1, 2) = '23' or
substr(c.c_phone, 1, 2) = '29' or
substr(c.c_phone, 1, 2) = '30' or
substr(c.c_phone, 1, 2) = '18' or
substr(c.c_phone, 1, 2) = '17')
group by
substring(c_phone, 1, 2)
order by
cntrycode
limit 100
---- PLAN
06:TOP-N [LIMIT=100]
| order by: substring(c_phone, 1, 2) ASC
|
05:AGGREGATE [FINALIZE]
| output: count(*), sum(c_acctbal)
| group by: substring(c_phone, 1, 2)
|
04:HASH JOIN [LEFT OUTER JOIN]
| hash predicates: c.c_custkey = o.o_custkey
| other predicates: o_custkey IS NULL
|
|--02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB compact
|
03:HASH JOIN [INNER JOIN]
| hash predicates: substr(c.c_name, 1, 1) = ct.cust_name_char
| other predicates: c.c_acctbal > ct.avg_acctbal
|
|--01:SCAN HDFS [tpch.q22_customer_tmp1 ct]
| partitions=1/1 size=30B compact
|
00:SCAN HDFS [tpch.customer c]
partitions=1/1 size=23.08MB
predicates: (substr(c.c_phone, 1, 2) = '13' OR substr(c.c_phone, 1, 2) = '31' OR substr(c.c_phone, 1, 2) = '23' OR substr(c.c_phone, 1, 2) = '29' OR substr(c.c_phone, 1, 2) = '30' OR substr(c.c_phone, 1, 2) = '18' OR substr(c.c_phone, 1, 2) = '17')
---- DISTRIBUTEDPLAN
12:MERGING-EXCHANGE [UNPARTITIONED]
| order by: substring(c_phone, 1, 2) ASC
| limit: 100
|
06:TOP-N [LIMIT=100]
| order by: substring(c_phone, 1, 2) ASC
|
11:AGGREGATE [MERGE FINALIZE]
| output: sum(count(*)), sum(sum(c_acctbal))
| group by: substring(c_phone, 1, 2)
|
10:EXCHANGE [HASH(substring(c_phone, 1, 2))]
|
05:AGGREGATE
| output: count(*), sum(c_acctbal)
| group by: substring(c_phone, 1, 2)
|
04:HASH JOIN [LEFT OUTER JOIN, PARTITIONED]
| hash predicates: c.c_custkey = o.o_custkey
| other predicates: o_custkey IS NULL
|
|--09:EXCHANGE [HASH(o.o_custkey)]
| |
| 02:SCAN HDFS [tpch.orders o]
| partitions=1/1 size=162.56MB
|
08:EXCHANGE [HASH(c.c_custkey)]
|
03:HASH JOIN [INNER JOIN, BROADCAST]
| hash predicates: substr(c.c_name, 1, 1) = ct.cust_name_char
| other predicates: c.c_acctbal > ct.avg_acctbal
|
|--07:EXCHANGE [BROADCAST]
| |
| 01:SCAN HDFS [tpch.q22_customer_tmp1 ct]
| partitions=1/1 size=30B
|
00:SCAN HDFS [tpch.customer c]
partitions=1/1 size=23.08MB
predicates: (substr(c.c_phone, 1, 2) = '13' OR substr(c.c_phone, 1, 2) = '31' OR substr(c.c_phone, 1, 2) = '23' OR substr(c.c_phone, 1, 2) = '29' OR substr(c.c_phone, 1, 2) = '30' OR substr(c.c_phone, 1, 2) = '18' OR substr(c.c_phone, 1, 2) = '17')
====
Reference in New Issue View Git Blame Copy Permalink