[
https://issues.apache.org/jira/browse/KYLIN-5038?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
star_dev updated KYLIN-5038:
----------------------------
Description:
整体验证思路为
1. 写case语句，查看相应的执行计划
2. case语句执行，看kylin.log找到相应的查询总结
3. 分析，改写case语句
case 1 临时表处理时间变化，join 临时表，where条件中的etl_dt与临时表日期进行判断
------------------------ SQL ------------------------
aaa as (
select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
'-', '' ) as sdate,
REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
as edate
)
SELECT
ETL_DT "日期",
sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
FROM
(
SELECT
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG,
count( 1 ) AS people,
sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
FROM
VIEW_SUBSCRIBER join aaa
on 1= 1
WHERE 1=1
and ETL_DT >= aaa.sdate
AND ETL_DT <= aaa.edate
AND COMPANY_ID = 17
AND LOCAL_ACTIVE_DAY <= '20210514'
AND WCT_FLAG = 0
AND T1_FLAG = 0
AND NDS_FLAG = 0
AND BUSINESS_ID != 3
AND DTH_FLAG = 1
AND DTV_FLAG = 0
GROUP BY
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG
)
GROUP BY
ETL_DT
ORDER BY
ETL_DT ASC
------------------------ SQL 执行计划------------------------
OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
充值金额（税前本地币）=[SUM($3)], ctx=[])
OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
$f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
PAY_MONEY_LOCAL=[SUM($3)], ctx=[])
OLAPProjectRel(ETL_DT=[$33], DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28],
PAY_MONEY_LOCAL=[$34], ctx=[])
OLAPFilterRel(condition=
[AND(=(1, 1),
>=($33, $52),
<=($33, $53),
=($3, 17),
<=($32, '20210514'),
=(CAST($30):INTEGER, 0),
=(CAST($29):INTEGER, 0),
=(CAST($31):INTEGER, 0),
<>($10, 3),
=(CAST($19):INTEGER, 1),
=(CAST($20):INTEGER, 0))], ctx=[])
OLAPJoinRel(
condition=[true],
joinType=[inner],
ctx=[])
OLAPTableScan(
table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2, 3, 4, 5, 6, 7,
8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27,
28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47,
48, 49, 50, 51]])
OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
-31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
NULL, '-', '')],
EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000, -1))):VARCHAR
CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT NULL, '-', '')],
ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
------------------------ SQL 查询总结------------------------
Success: true
Duration: 0.816
Project: dvb_ods
Realization Names: [CUBE[name=view_subscriber]]
Cuboid Ids: [8586246208]
Total scan count: 720352
Total scan bytes: 89324514
Result row count: 30
Accept Partial: true
Is Partial Result: false
Hit Exception Cache: false
Storage cache used: false
Is Query Push-Down: false
Is Prepare: false
Trace URL: null
Message: null
case 2 临时表处理时间变化，join 临时表，on条件中的etl_dt与临时表日期进行判断
------------------------ SQL ------------------------
aaa as (
select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
'-', '' ) as sdate,
REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
as edate
)
SELECT
ETL_DT "日期",
sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
FROM
(
SELECT
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG,
count( 1 ) AS people,
sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
FROM
VIEW_SUBSCRIBER join aaa
on ETL_DT >= aaa.sdate AND ETL_DT <= aaa.edate
WHERE 1=1
AND COMPANY_ID = 17
AND LOCAL_ACTIVE_DAY <= '20210514'
AND WCT_FLAG = 0
AND T1_FLAG = 0
AND NDS_FLAG = 0
AND BUSINESS_ID != 3
AND DTH_FLAG = 1
AND DTV_FLAG = 0
GROUP BY
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG
)
GROUP BY
ETL_DT
ORDER BY
ETL_DT ASC
------------------------ SQL 执行计划------------------------
OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
充值金额（税前本地币）=[SUM($3)], ctx=[])
OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
$f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
PAY_MONEY_LOCAL=[SUM($3)], ctx=[])
OLAPProjectRel(ETL_DT=[$33], DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28],
PAY_MONEY_LOCAL=[$34], ctx=[])
OLAPFilterRel(condition=[AND(>=($33, $52), <=($33, $53), =($3, 17), <=($32,
'20210514'), =(CAST($30):INTEGER, 0), =(CAST($29):INTEGER, 0),
=(CAST($31):INTEGER, 0), <>($10, 3), =(CAST($19):INTEGER, 1),
=(CAST($20):INTEGER, 0))], ctx=[])
OLAPJoinRel(condition=[true], joinType=[inner], ctx=[])
OLAPTableScan(table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2,
3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43,
44, 45, 46, 47, 48, 49, 50, 51]])
OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
-31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
NULL, '-', '')],
EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000, -1))):VARCHAR
CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT NULL, '-', '')],
ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
------------------------ SQL 查询总结------------------------
Success: true
Duration: 4.993
Project: dvb_ods
Realization Names: [CUBE[name=view_subscriber]]
Cuboid Ids: [8586246208]
Total scan count: 720352
Total scan bytes: 89324514
Result row count: 30
Accept Partial: true
Is Partial Result: false
Hit Exception Cache: false
Storage cache used: false
Is Query Push-Down: false
Is Prepare: false
Trace URL: null
Message: null
case 3 临时表处理时间变化，join 临时表，但是where的etl_dt直接写死日期，没有扫全表
------------------------ SQL ------------------------
aaa as (
select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
'-', '' ) as sdate,
REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
as edate
)
SELECT
ETL_DT "日期",
sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
FROM
(
SELECT
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG,
count( 1 ) AS people,
sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
FROM
VIEW_SUBSCRIBER join aaa
on 1= 1
WHERE 1=1
and ETL_DT >= '20210615'
AND ETL_DT <= '20210715'
AND COMPANY_ID = 17
AND LOCAL_ACTIVE_DAY <= '20210514'
AND WCT_FLAG = 0
AND T1_FLAG = 0
AND NDS_FLAG = 0
AND BUSINESS_ID != 3
AND DTH_FLAG = 1
AND DTV_FLAG = 0
GROUP BY
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG
)
GROUP BY
ETL_DT
ORDER BY
ETL_DT ASC
------------------------ SQL 执行计划------------------------
OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
充值金额（税前本地币）=[SUM($3)], ctx=[])
OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
$f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
PAY_MONEY_LOCAL=[SUM($3)], ctx=[]) OLAPProjectRel(ETL_DT=[$33],
DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28], PAY_MONEY_LOCAL=[$34], ctx=[])
OLAPFilterRel(condition=[AND(=(1, 1), >=($33, '20210615'), <=($33, '20210715'),
=($3, 17), <=($32, '20210514'), =(CAST($30):INTEGER, 0), =(CAST($29):INTEGER,
0), =(CAST($31):INTEGER, 0), <>($10, 3), =(CAST($19):INTEGER, 1),
=(CAST($20):INTEGER, 0))], ctx=[])
OLAPJoinRel(condition=[true], joinType=[inner], ctx=[])
OLAPTableScan(table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2, 3,
4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24,
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44,
45, 46, 47, 48, 49, 50, 51]])
OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
-31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
NULL, '-', '')], EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
-1))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
NULL, '-', '')], ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
------------------------ SQL 查询总结------------------------
Success: true
Duration: 4.712
Project: dvb_ods
Realization Names: [CUBE[name=view_subscriber]]
Cuboid Ids: [8586246208]
Total scan count: 561581
Total scan bytes: 69636323
Result row count: 30
Accept Partial: true
Is Partial Result: false
Hit Exception Cache: false
Storage cache used: false
Is Query Push-Down: false
Is Prepare: false
Trace URL: null
Message: null
case 4 临时表时间直接写死，join 临时表，where条件中的etl_dt与临时表日期进行判断
------------------------ SQL ------------------------
SQL: with abb as
(
select '20210715' as end_date,'20210615' as start_date )
select
ETL_DT "日期",
sum(
case
DEDUCTION_FLAG
when 1 THEN people
end
) "扣费用户数",
sum(
case
RECHARGE_FLAG
when 1 THEN people
end
) as "充值用户数",
sum(PAY_MONEY_LOCAL) "充值金额（税前本地币）"
from
(
SELECT
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG,
count(1) as people,
sum(PAY_MONEY_LOCAL) PAY_MONEY_LOCAL
from
VIEW_SUBSCRIBER
join abb on 1=1
where
ETL_DT >= start_date
and ETL_DT <= end_date
and COMPANY_ID = 17
and LOCAL_ACTIVE_DAY <= '20210514'
and WCT_FLAG = 0
and T1_FLAG = 0
and NDS_FLAG = 0
and BUSINESS_ID != 3
and DTH_FLAG = 1
and DTV_FLAG = 0
GROUP by
ETL_DT,
DEDUCTION_FLAG,
RECHARGE_FLAG
)
GROUP by
ETL_DT
ORDER by
ETL_DT asc
------------------------ SQL 执行计划------------------------
------------------------ SQL 查询总结------------------------
Success: true
Duration: 0.699
Project: dvb_ods
Realization Names: [CUBE[name=view_subscriber]]
Cuboid Ids: [8586246208]
Total scan count: 720352
Total scan bytes: 89324514
Result row count: 30
Accept Partial: true
Is Partial Result: false
Hit Exception Cache: false
Storage cache used: false
Is Query Push-Down: false
Is Prepare: false
Trace URL: null
Message: null
> Kylin每日增量构建（etl_dt），当分区字段etl_dt的条件值非固定值（如20210101），而是变量值时（如REPLACE ( cast(
> TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ), '-', '' )），扫全表
> ----------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: KYLIN-5038
> URL: https://issues.apache.org/jira/browse/KYLIN-5038
> Project: Kylin
> Issue Type: Bug
> Reporter: star_dev
> Priority: Major
>
> 整体验证思路为
> 1. 写case语句，查看相应的执行计划
> 2. case语句执行，看kylin.log找到相应的查询总结
> 3. 分析，改写case语句
>
> case 1 临时表处理时间变化，join 临时表，where条件中的etl_dt与临时表日期进行判断
> ------------------------ SQL ------------------------
> aaa as (
> select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
> '-', '' ) as sdate,
> REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
> as edate
> )
> SELECT
> ETL_DT "日期",
> sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
> sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
> sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
> FROM
> (
> SELECT
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG,
> count( 1 ) AS people,
> sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
> FROM
> VIEW_SUBSCRIBER join aaa
> on 1= 1
> WHERE 1=1
> and ETL_DT >= aaa.sdate
> AND ETL_DT <= aaa.edate
> AND COMPANY_ID = 17
> AND LOCAL_ACTIVE_DAY <= '20210514'
> AND WCT_FLAG = 0
> AND T1_FLAG = 0
> AND NDS_FLAG = 0
> AND BUSINESS_ID != 3
> AND DTH_FLAG = 1
> AND DTV_FLAG = 0
> GROUP BY
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG
> )
> GROUP BY
> ETL_DT
> ORDER BY
> ETL_DT ASC
> ------------------------ SQL 执行计划------------------------
> OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
> OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
> OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
> 充值金额（税前本地币）=[SUM($3)], ctx=[])
> OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
> $f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
> OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
> PAY_MONEY_LOCAL=[SUM($3)], ctx=[])
> OLAPProjectRel(ETL_DT=[$33], DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28],
> PAY_MONEY_LOCAL=[$34], ctx=[])
> OLAPFilterRel(condition=
> [AND(=(1, 1),
> >=($33, $52),
> <=($33, $53),
> =($3, 17),
> <=($32, '20210514'),
> =(CAST($30):INTEGER, 0),
> =(CAST($29):INTEGER, 0),
> =(CAST($31):INTEGER, 0),
> <>($10, 3),
> =(CAST($19):INTEGER, 1),
> =(CAST($20):INTEGER, 0))], ctx=[])
> OLAPJoinRel(
> condition=[true],
> joinType=[inner],
> ctx=[])
> OLAPTableScan(
> table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2, 3, 4, 5, 6,
> 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26,
> 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45,
> 46, 47, 48, 49, 50, 51]])
> OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
> -31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
> NULL, '-', '')],
> EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000, -1))):VARCHAR
> CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT NULL, '-',
> '')], ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
> ------------------------ SQL 查询总结------------------------
> Success: true
> Duration: 0.816
> Project: dvb_ods
> Realization Names: [CUBE[name=view_subscriber]]
> Cuboid Ids: [8586246208]
> Total scan count: 720352
> Total scan bytes: 89324514
> Result row count: 30
> Accept Partial: true
> Is Partial Result: false
> Hit Exception Cache: false
> Storage cache used: false
> Is Query Push-Down: false
> Is Prepare: false
> Trace URL: null
> Message: null
>
>
> case 2 临时表处理时间变化，join 临时表，on条件中的etl_dt与临时表日期进行判断
> ------------------------ SQL ------------------------
> aaa as (
> select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
> '-', '' ) as sdate,
> REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
> as edate
> )
> SELECT
> ETL_DT "日期",
> sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
> sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
> sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
> FROM
> (
> SELECT
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG,
> count( 1 ) AS people,
> sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
> FROM
> VIEW_SUBSCRIBER join aaa
> on ETL_DT >= aaa.sdate AND ETL_DT <= aaa.edate
> WHERE 1=1
> AND COMPANY_ID = 17
> AND LOCAL_ACTIVE_DAY <= '20210514'
> AND WCT_FLAG = 0
> AND T1_FLAG = 0
> AND NDS_FLAG = 0
> AND BUSINESS_ID != 3
> AND DTH_FLAG = 1
> AND DTV_FLAG = 0
> GROUP BY
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG
> )
> GROUP BY
> ETL_DT
> ORDER BY
> ETL_DT ASC
> ------------------------ SQL 执行计划------------------------
> OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
> OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
> OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
> 充值金额（税前本地币）=[SUM($3)], ctx=[])
> OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
> $f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
> OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
> PAY_MONEY_LOCAL=[SUM($3)], ctx=[])
> OLAPProjectRel(ETL_DT=[$33], DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28],
> PAY_MONEY_LOCAL=[$34], ctx=[])
> OLAPFilterRel(condition=[AND(>=($33, $52), <=($33, $53), =($3, 17), <=($32,
> '20210514'), =(CAST($30):INTEGER, 0), =(CAST($29):INTEGER, 0),
> =(CAST($31):INTEGER, 0), <>($10, 3), =(CAST($19):INTEGER, 1),
> =(CAST($20):INTEGER, 0))], ctx=[])
> OLAPJoinRel(condition=[true], joinType=[inner], ctx=[])
> OLAPTableScan(table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2,
> 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
> 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42,
> 43, 44, 45, 46, 47, 48, 49, 50, 51]])
> OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
> -31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
> NULL, '-', '')],
> EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000, -1))):VARCHAR
> CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT NULL, '-',
> '')], ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
> ------------------------ SQL 查询总结------------------------
> Success: true
> Duration: 4.993
> Project: dvb_ods
> Realization Names: [CUBE[name=view_subscriber]]
> Cuboid Ids: [8586246208]
> Total scan count: 720352
> Total scan bytes: 89324514
> Result row count: 30
> Accept Partial: true
> Is Partial Result: false
> Hit Exception Cache: false
> Storage cache used: false
> Is Query Push-Down: false
> Is Prepare: false
> Trace URL: null
> Message: null
>
>
> case 3 临时表处理时间变化，join 临时表，但是where的etl_dt直接写死日期，没有扫全表
> ------------------------ SQL ------------------------
> aaa as (
> select REPLACE ( cast( TIMESTAMPADD( DAY,- 31, CURRENT_DATE ) AS VARCHAR ),
> '-', '' ) as sdate,
> REPLACE ( cast( TIMESTAMPADD( DAY,- 1, CURRENT_DATE ) AS VARCHAR ), '-', '' )
> as edate
> )
> SELECT
> ETL_DT "日期",
> sum( CASE DEDUCTION_FLAG WHEN 1 THEN people END ) "扣费用户数",
> sum( CASE RECHARGE_FLAG WHEN 1 THEN people END ) AS "充值用户数",
> sum( PAY_MONEY_LOCAL ) "充值金额（税前本地币）"
> FROM
> (
> SELECT
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG,
> count( 1 ) AS people,
> sum( PAY_MONEY_LOCAL ) PAY_MONEY_LOCAL
> FROM
> VIEW_SUBSCRIBER join aaa
> on 1= 1
> WHERE 1=1
> and ETL_DT >= '20210615'
> AND ETL_DT <= '20210715'
> AND COMPANY_ID = 17
> AND LOCAL_ACTIVE_DAY <= '20210514'
> AND WCT_FLAG = 0
> AND T1_FLAG = 0
> AND NDS_FLAG = 0
> AND BUSINESS_ID != 3
> AND DTH_FLAG = 1
> AND DTV_FLAG = 0
> GROUP BY
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG
> )
> GROUP BY
> ETL_DT
> ORDER BY
> ETL_DT ASC
> ------------------------ SQL 执行计划------------------------
> OLAPToEnumerableConverter OLAPLimitRel(ctx=[], fetch=[50000])
> OLAPSortRel(sort0=[$0], dir0=[ASC], ctx=[])
> OLAPAggregateRel(group=[\{0}], 扣费用户数=[SUM($1)], 充值用户数=[SUM($2)],
> 充值金额（税前本地币）=[SUM($3)], ctx=[])
> OLAPProjectRel(日期=[$0], $f1=[CASE(=($1, 1), CAST($3):BIGINT, null)],
> $f2=[CASE(=($2, 1), CAST($3):BIGINT, null)], PAY_MONEY_LOCAL=[$4], ctx=[])
> OLAPAggregateRel(group=[\{0, 1, 2}], PEOPLE=[COUNT()],
> PAY_MONEY_LOCAL=[SUM($3)], ctx=[]) OLAPProjectRel(ETL_DT=[$33],
> DEDUCTION_FLAG=[$27], RECHARGE_FLAG=[$28], PAY_MONEY_LOCAL=[$34], ctx=[])
> OLAPFilterRel(condition=[AND(=(1, 1), >=($33, '20210615'), <=($33,
> '20210715'), =($3, 17), <=($32, '20210514'), =(CAST($30):INTEGER, 0),
> =(CAST($29):INTEGER, 0), =(CAST($31):INTEGER, 0), <>($10, 3),
> =(CAST($19):INTEGER, 1), =(CAST($20):INTEGER, 0))], ctx=[])
> OLAPJoinRel(condition=[true], joinType=[inner], ctx=[])
> OLAPTableScan(table=[[BOSS_ODS, VIEW_SUBSCRIBER]], ctx=[], fields=[[0, 1, 2,
> 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23,
> 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42,
> 43, 44, 45, 46, 47, 48, 49, 50, 51]])
> OLAPProjectRel(SDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
> -31))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
> NULL, '-', '')], EDATE=[REPLACE(CAST(DATETIME_PLUS(CURRENT_DATE, *(86400000,
> -1))):VARCHAR CHARACTER SET "UTF-16LE" COLLATE "UTF-16LE$en_US$primary" NOT
> NULL, '-', '')], ctx=[]) OLAPValuesRel(tuples=[[\{ 0 }]])
> ------------------------ SQL 查询总结------------------------
> Success: true
> Duration: 4.712
> Project: dvb_ods
> Realization Names: [CUBE[name=view_subscriber]]
> Cuboid Ids: [8586246208]
> Total scan count: 561581
> Total scan bytes: 69636323
> Result row count: 30
> Accept Partial: true
> Is Partial Result: false
> Hit Exception Cache: false
> Storage cache used: false
> Is Query Push-Down: false
> Is Prepare: false
> Trace URL: null
> Message: null
>
>
> case 4 临时表时间直接写死，join 临时表，where条件中的etl_dt与临时表日期进行判断
> ------------------------ SQL ------------------------
> SQL: with abb as
> (
> select '20210715' as end_date,'20210615' as start_date )
>
> select
> ETL_DT "日期",
> sum(
> case
> DEDUCTION_FLAG
> when 1 THEN people
> end
> ) "扣费用户数",
> sum(
> case
> RECHARGE_FLAG
> when 1 THEN people
> end
> ) as "充值用户数",
> sum(PAY_MONEY_LOCAL) "充值金额（税前本地币）"
> from
> (
> SELECT
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG,
> count(1) as people,
> sum(PAY_MONEY_LOCAL) PAY_MONEY_LOCAL
> from
> VIEW_SUBSCRIBER
> join abb on 1=1
> where
> ETL_DT >= start_date
> and ETL_DT <= end_date
>
> and COMPANY_ID = 17
> and LOCAL_ACTIVE_DAY <= '20210514'
> and WCT_FLAG = 0
> and T1_FLAG = 0
> and NDS_FLAG = 0
> and BUSINESS_ID != 3
> and DTH_FLAG = 1
> and DTV_FLAG = 0
> GROUP by
> ETL_DT,
> DEDUCTION_FLAG,
> RECHARGE_FLAG
> )
> GROUP by
> ETL_DT
> ORDER by
> ETL_DT asc
> ------------------------ SQL 执行计划------------------------
> ------------------------ SQL 查询总结------------------------
> Success: true
> Duration: 0.699
> Project: dvb_ods
> Realization Names: [CUBE[name=view_subscriber]]
> Cuboid Ids: [8586246208]
> Total scan count: 720352
> Total scan bytes: 89324514
> Result row count: 30
> Accept Partial: true
> Is Partial Result: false
> Hit Exception Cache: false
> Storage cache used: false
> Is Query Push-Down: false
> Is Prepare: false
> Trace URL: null
> Message: null
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
