感谢您的帮助/建议。我不清楚LAG
表达式中的1,0,那是什么?为什么我的代码不起作用?
我是否必须对超前和滞后都做两个Order by
?
Select
* Customer,
Prod,
day,
current sold,
date,
lag[current sold,1,0] OVER(PARTITION BY customer ORDER BY date DESC) as Previous Day,
lead[current sold,1,0] OVER(PARTITION BY customer ORDER BY date DESC) as Next Day
From
table1
结果:
| PROD | DAY | CURRENT SOLD | date customer |
+-------+-----+--------------+-----------------------
| SHIRT | M | 2 | 1-2018 A
| SHIRT | T | 9 | 2-2018 B
| SHIRT | W | 0 | 12-2018 C
| SHIRT | TH | 6 | 11-2018 D
| SHIRT | F | 7 | 3-2018 E
+-------+-----+--------------+--+----------------
+-------+-----+--------------+---------------+-----------+--+--------------
| PROD | DAY | CURRENT SOLD | PREVIOUS SOLD | NEXT SOLD | date |customer
+-------+-----+--------------+---------------+-----------+--+---------------
| SHIRT | M | 2 | | 9 | 1-2018 |A
| SHIRT | T | 9 | 2 | 0 | 2-2018 |B
| SHIRT | W | 0 | 9 | 6 | 12-2018|C
| SHIRT | TH | 6 | 0 | 7 | 11-2018|D
| SHIRT | F | 7 | 6 | | 3-2018 |E
+-------+-----+--------------+---------------+-----------+--+---------------
答案 0 :(得分:0)
您可以将LAG用于上一次销售,而LEAD用于下一次销售。我为您的示例准备了样品:
--DROP TABLE #Tbl;
--DROP TABLE #Days;
CREATE TABLE #Tbl
(
Prod VARCHAR(10)
,[DayName] VARCHAR(2)
,CurrentSold INT
);
CREATE TABLE #Days
(
DayNumber INT
,[DayName] VARCHAR(2)
);
INSERT INTO #Days
VALUES (1,'M'),(2,'T'),(3,'W'),(4,'TH'),(5,'F');
INSERT INTO #Tbl
VALUES ('SHIRT','M',2)
,('SHIRT','T',9)
,('SHIRT','W',0)
,('SHIRT','TH',6)
,('SHIRT','F',7);
SELECT T.Prod
,T.DayName
,T.CurrentSold
,LAG(CurrentSold, 1,0) OVER (ORDER BY DayNumber) AS PreviousSold
,LEAD(CurrentSold, 1,0) OVER (ORDER BY DayNumber) AS PreviousSold
FROM #Tbl T
INNER JOIN #Days D ON T.DayName = D.DayName;