使用维度表来处理主表中不存在的数据组合。
问题描述 投票：0回答：1
我有一个生成报表的存储过程。实际的报告有点复杂，所以我将尽量用表格的例子来简单解释。
我的主表有以下数据。
我的主表有以下数据： table1
╔════╦═══════╦═════════╦═══════╗
║ id ║ price ║ Item_id ║ name ║
╠════╬═══════╬═════════╬═══════╣
║ 1 ║ 3 ║ 11 ║ item1 ║
╠════╬═══════╬═════════╬═══════╣
║ 2 ║ 5 ║ 22 ║ item2 ║
╠════╬═══════╬═════════╬═══════╣
║ 3 ║ 6 ║ 33 ║ item3 ║
╠════╬═══════╬═════════╬═══════╣
║ 4 ║ 7 ║ 44 ║ item4 ║
╠════╬═══════╬═════════╬═══════╣
║ 5 ║ 8 ║ 55 ║ item5 ║
╚════╩═══════╩═════════╩═══════╝
我有代码。
select id,
CASE
when price <= 2 THEN '01-02'
when price <= 3 THEN '03-04'
when price <= 4 THEN '05-05'
when price <= 5 THEN '05-06'
...ETC
END as price,
Item_id,
name
from table1
现在，从我的数据来看，在这种情况下，没有价格"。when price <= 2 THEN '01-02'"所以该选项不会出现在最终的报告中，然而我希望它能出现，用0。
我明白，为此我需要用 "维度 "表做一个LEFT JOIN来使用模板，但我不知道如何实现。
我不知道会不会是一个表，里面有我想显示的所有组合，比如说像 "'01-02'" 然后加入它，但这似乎并不奏效。
这是SP的代码。
DECLARE @RaceTrack CHAR(5)= 'Bala';
DECLARE @Category VARCHAR(4000)= 'All, Class, Days, Distance, Race, Runners, Total Bracket Count, Runners2';
DECLARE @TrotPace VARCHAR(10)= 'Excluded';
WITH CTE_TblData
AS (SELECT td.EVENT_ID,
td.MENU_HINT,
SUBSTRING(td.menu_hint, 7, 4) AS [RaceTrack],
td.EVENT_NAME,
td.BSP,
td.EVENT_DT AS Expr1003,
td.WIN_LOSE,
COUNT(*) OVER(PARTITION BY event_id) AS [RaceRunners],
td.[Race No] AS RaceNo,
replace(LEFT(SUBSTRING(td.EVENT_NAME, CHARINDEX(' ', td.EVENT_NAME) + 1, LEN(td.EVENT_NAME)), CHARINDEX(' ', SUBSTRING(td.EVENT_NAME, CHARINDEX(' ', td.EVENT_NAME) + 2, LEN(td.EVENT_NAME)))), 'm', '') AS Distance,
REVERSE(LEFT(REVERSE(td.EVENT_NAME), CHARINDEX(' ', REVERSE(td.EVENT_NAME)) - 1)) AS Class,
FORMAT(CONVERT(DATETIME, td.EVENT_DT, 103), 'dddd') AS [Day],
DATEPART(WEEKDAY, CONVERT(DATETIME, td.EVENT_DT, 103)) AS [DayNumber]
FROM tblData td
WHERE(@TrotPace <> 'Excluded' OR td.event_name NOT LIKE '%trot%' AND td.event_name NOT LIKE '%pace%') AND SUBSTRING(td.menu_hint, 7, 4) = @RaceTrack),
NewCTE1_PriceLabel
AS (SELECT td.*, price.label AS [PriceBracketLabel]
FROM CTE_TblData AS td
CROSS APPLY
(
SELECT CASE WHEN td.BSP <= 2 THEN '01-02'
WHEN td.BSP <= 3 THEN '02-03'
WHEN td.BSP <= 5 THEN '03-05'
WHEN td.BSP <= 8 THEN '05-08'
WHEN td.BSP <= 13 THEN '08-13'
WHEN td.BSP <= 21 THEN '13-21'
WHEN td.BSP <= 34 THEN '21-34'
WHEN td.BSP <= 55 THEN '34-55'
WHEN td.BSP <= 89 THEN '55-89'
WHEN td.BSP <= 1000 THEN '89-1000'
END AS label
) AS price),
NewCTE2_WindowedCount
AS (SELECT *,
COUNT(*) OVER(PARTITION BY event_id,
PriceBracketLabel) AS [PriceBracketRunners]
FROM NewCTE1_PriceLabel),
DataCube
AS (SELECT td.event_id,
td.MENU_HINT,
td.EVENT_NAME,
td.Expr1003,
td.RaceNo,
td.Distance,
td.Class,
td.[Day],
CASE WHEN brackets.category = brackets2.category then brackets.category ELSE brackets.category + ' : ' + brackets2.category END AS bracket_category,
CASE WHEN brackets.label = brackets2.label THEN td.PriceBracketlabel + ' : ' + brackets.label ELSE td.PriceBracketlabel + ' : ' + brackets.label + ' : ' + brackets2.label END AS bracket_label,
SUM(CASE WHEN td.WIN_LOSE = 1 THEN 1 ELSE 0 END) AS CountHorses_Winners,
SUM(CASE WHEN td.WIN_LOSE = 1 THEN td.BSP ELSE 0 END) AS SumPrices_Winners,
SUM(CASE WHEN td.WIN_LOSE = 0 THEN 1 ELSE 0 END) AS CountHorses_Losers,
SUM(CASE WHEN td.WIN_LOSE = 0 THEN td.BSP ELSE 0 END) AS SumPrices_Losers
FROM NewCTE2_WindowedCount AS td
CROSS APPLY
(
SELECT 'All' AS category,
'All' AS label
UNION ALL
SELECT 'Runners' AS category,
CASE
WHEN td.RaceRunners <= 7 THEN '01-07'
WHEN td.RaceRunners <= 12 THEN '07-12'
WHEN td.RaceRunners <= 1000 THEN '12-1000'
END AS label
UNION ALL
SELECT 'Total Bracket Count' AS category,
CASE
WHEN td.PriceBracketRunners = 1 THEN '01'
WHEN td.PriceBracketRunners IN(2, 3) THEN '02-03'
WHEN td.PriceBracketRunners > 3 THEN '04-1000'
END AS label
UNION ALL
SELECT 'Distance' AS category,
CASE
WHEN td.distance <= 1200 THEN '0-1200'
WHEN td.distance <= 1600 THEN '1200-1600'
WHEN td.distance <= 10000 THEN '1600-10000'
END AS label
UNION ALL
SELECT 'Race' AS category,
CASE
WHEN td.raceno < 5 THEN '0-4'
WHEN td.raceno < 1000 THEN '04-1000'
END AS label
UNION ALL
--the stuff after the "then" is the title for the column, the stuff inside the brackets of "in" is what it filters
SELECT 'Class' AS category,
CASE
WHEN td.Class IN('Mdn') THEN 'Mdn'
WHEN td.Class IN('Cl1', 'Cl2', 'Cl3') THEN 'Cl1,Cl2,Cl3'
WHEN td.Class IN('Cl4', 'Cl5', 'Cl6') THEN 'Cl4,Cl5,Cl6'
WHEN td.Class IN('2yo') THEN '2yo'
WHEN td.Class IN('3yo') THEN '3yo'
WHEN td.Class IN('Hcp', 'Hcap') THEN 'Hcp'
WHEN td.Class IN('Qlty', 'Listed', 'Grp3', 'Grp2', 'Grp1', 'WFA') THEN 'Qlty,Listed,Grp3,Grp2,Grp1,WFA'
END AS label
UNION ALL
SELECT 'Days' AS category,
CASE
WHEN td.DayNumber IN(1, 2, 3) THEN 'Sun-Mon-Tues'
WHEN td.DayNumber IN(4, 5) THEN 'Wed-Sat'
WHEN td.DayNumber IN(6, 7) THEN 'Thurs-Fri'
END AS label
) AS brackets
cross apply
(
SELECT 'All' AS category,
'All' AS label
UNION ALL
SELECT 'Runners' AS category,
CASE
WHEN td.RaceRunners <= 7 THEN '01-07'
WHEN td.RaceRunners <= 12 THEN '07-12'
WHEN td.RaceRunners <= 1000 THEN '12-1000'
END AS label
UNION ALL
SELECT 'Total Bracket Count' AS category,
CASE
WHEN td.PriceBracketRunners = 1 THEN '01'
WHEN td.PriceBracketRunners IN(2, 3) THEN '02-03'
WHEN td.PriceBracketRunners > 3 THEN '04-1000'
END AS label
UNION ALL
SELECT 'Distance' AS category,
CASE
WHEN td.distance <= 1200 THEN '0-1200'
WHEN td.distance <= 1600 THEN '1200-1600'
WHEN td.distance <= 10000 THEN '1600-10000'
END AS label
UNION ALL
SELECT 'Race' AS category,
CASE
WHEN td.raceno < 5 THEN '0-4'
WHEN td.raceno < 1000 THEN '04-1000'
END AS label
UNION ALL
--the stuff after the "then" is the title for the column, the stuff inside the brackets of "in" is what it filters
SELECT 'Class' AS category,
CASE
WHEN td.Class IN('Mdn') THEN 'Mdn'
WHEN td.Class IN('Cl1', 'Cl2', 'Cl3') THEN 'Cl1,Cl2,Cl3'
WHEN td.Class IN('Cl4', 'Cl5', 'Cl6') THEN 'Cl4,Cl5,Cl6'
WHEN td.Class IN('2yo') THEN '2yo'
WHEN td.Class IN('3yo') THEN '3yo'
WHEN td.Class IN('Hcp', 'Hcap') THEN 'Hcp'
WHEN td.Class IN('Qlty', 'Listed', 'Grp3', 'Grp2', 'Grp1', 'WFA') THEN 'Qlty,Listed,Grp3,Grp2,Grp1,WFA'
END AS label
UNION ALL
SELECT 'Days' AS category,
CASE
WHEN td.DayNumber IN(1, 2, 3) THEN 'Sun-Mon-Tues'
WHEN td.DayNumber IN(4, 5) THEN 'Wed-Sat'
WHEN td.DayNumber IN(6, 7) THEN 'Thurs-Fri'
END AS label
) AS brackets2
GROUP BY td.event_id,
td.MENU_HINT,
td.EVENT_NAME,
td.Expr1003,
td.RaceNo,
td.Distance,
td.Class,
td.[Day],
brackets.category,
brackets2.category,
td.PriceBracketlabel,
brackets.label,
brackets2.label),
WinnerGroups
AS (SELECT *,
SUM(CountHorses_Winners) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY Expr1003,
RaceNo,
event_id) AS winner_group_id,
SumPrices_Winners - CountHorses_Winners - CountHorses_Losers AS line_profit
FROM DataCube),
GroupTotals
AS (SELECT bracket_category,
bracket_label,
winner_group_id,
SUM(line_profit) AS Group_Total_Profit
FROM WinnerGroups
GROUP BY bracket_category,
bracket_label,
winner_group_id),
GroupSteppedTotals
AS (SELECT bracket_category,
bracket_label,
winner_group_id,
Group_Total_Profit,
SUM(Group_Total_Profit) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY winner_group_id ROWS BETWEEN 0 PRECEDING AND CURRENT ROW) AS Group_Total1_Profit,
SUM(Group_Total_Profit) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY winner_group_id ROWS BETWEEN 1 PRECEDING AND CURRENT ROW) AS Group_Total2_Profit,
SUM(Group_Total_Profit) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY winner_group_id ROWS BETWEEN 2 PRECEDING AND CURRENT ROW) AS Group_Total3_Profit,
SUM(Group_Total_Profit) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY winner_group_id ROWS BETWEEN 4 PRECEDING AND CURRENT ROW) AS Group_Total5_Profit,
SUM(Group_Total_Profit) OVER(PARTITION BY bracket_category,
bracket_label
ORDER BY winner_group_id ROWS BETWEEN 7 PRECEDING AND CURRENT ROW) AS Group_Total8_Profit
FROM GroupTotals),
RunningTotals
AS (SELECT *,
SUM(line_profit) OVER(PARTITION BY bracket_category,
bracket_label,
winner_group_id
ORDER BY Expr1003 DESC,
RaceNo DESC,
event_id DESC) - line_profit AS Group_Partial_Profit
FROM WinnerGroups),
LaggedProfits
AS (SELECT RunningTotals.*,
GroupSteppedTotals.Group_Total_Profit,
LAG(RunningTotals.winner_group_id) OVER(PARTITION BY RunningTotals.bracket_category,
RunningTotals.bracket_label
ORDER BY RunningTotals.Expr1003,
RunningTotals.RaceNo,
RunningTotals.event_id) AS previous_winner_group_id,
(GroupSteppedTotals.Group_Total1_Profit - RunningTotals.Group_Partial_Profit) AS profit_back1winner,
(GroupSteppedTotals.Group_Total2_Profit - RunningTotals.Group_Partial_Profit) AS profit_back2winner,
(GroupSteppedTotals.Group_Total3_Profit - RunningTotals.Group_Partial_Profit) AS profit_back3winner,
(GroupSteppedTotals.Group_Total5_Profit - RunningTotals.Group_Partial_Profit) AS profit_back5winner,
(GroupSteppedTotals.Group_Total8_Profit - RunningTotals.Group_Partial_Profit) AS profit_back8winner
FROM RunningTotals
INNER JOIN GroupSteppedTotals ON GroupSteppedTotals.bracket_category = RunningTotals.bracket_category
AND GroupSteppedTotals.bracket_label = RunningTotals.bracket_label
AND GroupSteppedTotals.winner_group_id = RunningTotals.winner_group_id)
SELECT
event_id,
MENU_HINT,
EVENT_NAME,
Expr1003,
Class,
RaceNo,
lp.bracket_category,
bracket_label,
-- winner_group_id, -- Debugging Column
CountHorses_Winners,
SumPrices_Winners,
CountHorses_Losers,
SumPrices_Losers,
-- line_profit, -- Debugging Column
-- Group_Partial_Profit, -- Debugging Column
-- Group_Total_Profit, -- Debugging Column
CASE WHEN CountHorses_Winners > 0 AND previous_winner_group_id >= 1 THEN profit_back1winner END AS Profit_Winner_1,
CASE WHEN CountHorses_Winners > 0 AND previous_winner_group_id >= 2 THEN profit_back2winner END AS Profit_Winner_2,
CASE WHEN CountHorses_Winners > 0 AND previous_winner_group_id >= 3 THEN profit_back3winner END AS Profit_Winner_3,
CASE WHEN (previous_winner_group_id >= 3 and CountHorses_Winners <> 0) THEN cast(((profit_back1winner + profit_back2winner +profit_back3winner)/3) as decimal(10,2)) END AS Average_Winner,
-- CASE WHEN CountHorses_Winners > 0 AND previous_winner_group_id >= 5 THEN profit_back5winner END AS Profit_Winner_5,
-- CASE WHEN CountHorses_Winners > 0 AND previous_winner_group_id >= 8 THEN profit_back8winner END AS Profit_Winner_8,
CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 1 AND CountHorses_Losers > 0 THEN profit_back1winner END AS Profit_NotWinner_1,
CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 2 AND CountHorses_Losers > 0 THEN profit_back2winner END AS Profit_NotWinner_2,
CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 3 AND CountHorses_Losers > 0 THEN profit_back3winner END AS Profit_NotWinner_3,
CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 3 AND CountHorses_Losers > 0 THEN cast(((profit_back1winner + profit_back2winner +profit_back3winner)/3) as decimal(10,2)) END AS Average_NotWinner
-- CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 5 AND CountHorses_Losers > 0 THEN profit_back5winner END AS Profit_NotWinner_5,
-- CASE WHEN CountHorses_Winners = 0 AND previous_winner_group_id >= 8 AND CountHorses_Losers > 0 THEN profit_back8winner END AS Profit_NotWinner_8
FROM LaggedProfits lp
WHERE lp.bracket_category in ('All', 'Class', 'Days', 'Distance', 'Race', 'Runners', 'Total Bracket Count', 'Class : Days','Class : Distance','Class : Race', 'Class : Runners', 'Class : Total Bracket Count', 'Days : Distance', 'Days : Race', 'Days : Runners',
'Days : Total Bracket Count', 'Distance : Race', 'Distance : Runners', 'Distance : Total Bracket Count', 'Race : Runners', 'Race : Total Bracket Count', 'Runners : Total Bracket Count')
--and bracket_label = '03-05 : All'
-- WHERE --@Category IS NULL OR bracket_category IN (SELECT Value FROM dbo.FnSplit(@Category, ',')) and
--event_id = 145688261
--AND LaggedProfits.MENU_HINT LIKE '%MVal%'
--bracket_category = 'Runners2' AND
-- and bracket_label = '03-05 : Hcp : Wed-Sat'
--LaggedProfits.MENU_HINT LIKE '%Flem%'
ORDER BY Expr1003 DESC,
RaceNo DESC,
event_id DESC,
bracket_category,
bracket_label,
Class;
这是一个表格结果的例子，如果你看到，"01 -02 ... ...全部 "和 "02 -03 ... ...全部 "的 "Bracket_Label "没有出现。全部 "和 "02-03 : 全部 "没有出现。
1个回答
1
投票
投票
你可以使用 left join. 我会推荐。
select v.price, t1.Item_id, t1.name
from (values (0, 2, '01-02'),
(2, 4, '03-04'),
. . .
) v(lo, hi, price) left join
table1 t1
on t1.lo > v.lo and t1.hi <= v.hi;
最新问题
- beanstalk docker - 如何使用 Docker.run.json v1 设置容器名称
- 查找具有相同行数的两个二维数组之间的关联行差异
- decoder.ptr 为 nullptr CreateDesktopWindowTarget - IDesktopWindowTarget
- sh: 1: less: 使用PSQL命令时找不到
- 多维关联数组 - 差异
- 为什么在添加新项目后 React Native FlatList 没有更新？
- 如何在Android中加载Flutter资源
- 更改传递给 Python 函数的变量的内容。从变量创建引用的语法糖？
- SSRS - 导航到链接的网站或报告时，然后导航回来，丢失滚动位置和参数
- Azure AD Connect“无法验证凭据”
- 有什么方法可以通过CSS从SVG背景图像中删除/禁用feFilter吗？
- GCP Cloud Run 抱怨“无法打开文件 '/app/main.py'：[Errno 2] 没有这样的文件或目录”
- 下一个js应用程序路由器，如果已经登录则重定向到主页
- SSH和github如何确定用户名？
- Authjs 的 OAuth 的 Google 提供商无法在 Nextjs 中工作
- TizenBrew 错误，安装失败 检查证书错误::签名中包含证书的证书链无效。:<-3>
- 在groupByKey或groupBy之后形成列列表
- 检测字符串中的字符是否是表情符号（使用Android）
- 动态生成具有静态内容的函数
- 在 React 中从 AWS S3 预签名 url 下载文件
© www.soinside.com 2019 - 2024. All rights reserved.