给定一个任务、人员和时间表，我希望能够看到在每项任务上花费最多时间的人、他们在该任务上花费了多少时间以及下方的摘要行。
示例表：
task person time
---- ------ ----
Admin Sue 0.5
Admin Ted 0.25
Meetings Ted 1.25
Meetings Sue 0.75
期望的结果：
task top_person
---- ----------
Admin Sue (0.5)
Meetings Ted (1.25)
Overall Ted (1.5)
所以我们可以看到每个任务排名最高的人以及总体排名最高的人以及他们在所有任务中的总时间。
到目前为止我有什么：
SELECT
DECODE(GROUPING(task), 0, task, 'Overall') AS task,
MAX(person || ' (' || time || ')') KEEP (DENSE_RANK FIRST ORDER BY time DESC) AS top_person
FROM time_table
GROUP BY ROLLUP(task)
输出：
task top_person
---- ----------
Admin Sue (0.5)
Meetings Ted (1.25)
Overall Ted (1.25)
此查询按任务对表进行分组，按时间对组内的任务进行排序，保留第一行，然后从保留的行中显示人员及其时间。除了汇总行外，这行得通，似乎这条行对各个任务组执行相同的逻辑，但跨所有任务除外。相反，我需要的是所有任务的时间总和，按人分组，按小时排序，但仅适用于汇总行，而不适用于其他行。关于如何实现这一目标的任何想法？
也许您可以考虑对此类查询使用 MODEL 子句。行间/列间操作非常方便
示例数据：
WITH
tbl (TASK, PERSON, TASK_TIME) AS
(
Select 'Admin', 'Sue', 0.5 FRom Dual Union All
Select 'Admin', 'Ted', 0.25 FRom Dual Union All
Select 'Meetings', 'Ted', 1.25 FRom Dual Union All
Select 'Meetings', 'Sue', 0.75 FRom Dual
)
SQL 问题中的结果：
Select *
From ( Select TASK_STR "TASK", PERSON_STR "TOP_PERSON"
From ( Select CAST(TASK as Varchar2(20)) "TASK",
CAST(TASK as Varchar2(20)) "TASK_STR",
CAST(PERSON as VarChar2(20)) "PERSON",
TASK_TIME "TASK_TIME",
'(' || PERSON || ' ' || TASK_TIME || ')' "PERSON_STR"
From tbl
)
MODEL Dimension By (TASK, PERSON)
Measures(TASK_TIME, TASK_STR, PERSON_STR)
Rules (
TASK_STR[ANY, ANY] = Case When TASK_TIME[CV(), CV()] = Max(TASK_TIME)[CV(), ANY] Then CV(TASK) END,
TASK_STR['Overall', 'Time'] = CV(TASK) || ' ' || CV(PERSON),
PERSON_STR['Overall', 'Time'] = Sum(TASK_TIME)[TASK != CV(), PERSON != CV()]
)
)
Where TASK Is Not Null
R e s u l t :
TASK TOP_PERSON
------------------------------ ------------------------------------
Admin (Sue .5)
Meetings (Ted 1.25)
Overall Time 2.75
更改 MODEL 子句的 RULES 部分可以为您提供您可能需要或想从数据中获得的任何其他信息：
... ... ...
MODEL Dimension By (TASK, PERSON)
Measures(TASK_TIME, TASK_STR, PERSON_STR)
Rules (
TASK_STR[ANY, ANY] = Case When TASK_TIME[CV(), CV()] = Max(TASK_TIME)[CV(), ANY] Then CV(TASK) END,
--
TASK_STR['Overall', 'Data'] = 'Overall Data Below',
PERSON_STR['Overall', 'Data'] = LPAD('-', 15, '-'),
--
TASK_STR['Time', 'Total'] = CV(TASK) || ' ' || CV(PERSON),
PERSON_STR['Time', 'Total'] = Sum(TASK_TIME)[TASK != CV(), PERSON != CV()],
--
TASK_STR['Rows', 'Count'] = 'Rows Count',
PERSON_STR['Rows', 'Count'] = Count(PERSON_STR)[TASK != CV() And TASK NOT IN('Overall', 'Time', 'Rows'), PERSON != CV() And PERSON NOT IN('Data', 'Total', 'Count', 'Average')],
--
TASK_STR['Time', 'Average'] = 'Time Average',
PERSON_STR['Time', 'Average'] = Avg(TASK_TIME)[TASK != CV() And TASK != 'Overall', PERSON != CV() And PERSON NOT IN('Data', 'Total', 'Count', 'Average')]
)
)
Where TASK Is Not Null
Order By TASK
R e s u l t :
TASK TOP_PERSON
-------------------- ----------------------------------------------
Admin (Sue .5)
Meetings (Ted 1.25)
Overall Data Below ---------------
Rows Count 4
Time Average .6875
Time Total 2.75
虽然 Model 子句工作得非常好，但是如果你想要更多 SQL 特定的解决方案，你可以尝试 -
WITH max_data AS (SELECT task, person, time,
ROW_NUMBER() OVER(PARTITION BY task ORDER BY time DESC) RN1
FROM your_table
),
overall_person_data AS (SELECT person, SUM(time) sum_time
FROM your_table
GROUP BY person
)
SELECT task, person || '(' || max_time || ')'
FROM max_data
WHERE RN1 = 1
UNION ALL
SELECT 'Overall', person || '(' || MAX(sum_time) || ')'
FROM max_data
GROUP BY person;
更正了原始答案的语法
WITH
tbl (TASK, PERSON, TASK_TIME) AS
(
Select 'Admin', 'Sue', 0.5 FRom Dual Union All
Select 'Admin', 'Ted', 0.25 FRom Dual Union All
Select 'Meetings', 'Ted', 1.25 FRom Dual Union All
Select 'Meetings', 'Sue', 0.75 FRom Dual
),
max_data AS
(SELECT task, person, task_time,
ROW_NUMBER() OVER(PARTITION BY task ORDER BY task_time DESC) RN1
FROM tbl
),
overall_person_data AS
(SELECT person, SUM(task_time) sum_time
FROM max_data
GROUP BY person
)
Main SQL 1.
SELECT task, person || '(' || task_time || ')' "TOP_PERSON"
FROM max_data
WHERE RN1 = 1
UNION ALL
SELECT 'Overall', person || '(' || SUM(sum_time) || ')'
FROM overall_person_data
GROUP BY person
RESULT 1
TASK TOP_PERSON
-------- ---------------------------------------------
Admin Sue(.5)
Meetings Ted(1.25)
Overall Sue(1.25)
Overall Ted(1.5)
Main SQL 2.
SELECT task, person || '(' || task_time || ')' "TOP_PERSON"
FROM max_data
WHERE RN1 = 1
UNION ALL
SELECT 'Overall', '(' || SUM(sum_time) || ')'
FROM overall_person_data
RESULT 2.
TASK TOP_PERSON
-------- ---------------------------------------------
Admin Sue(.5)
Meetings Ted(1.25)
Overall (2.75)