Question

我正在尝试根据同一个表中不同列中的值更新sql server 2014中的表列。这是表格的一小部分。

CREATE TABLE [dbo].[Table1](
    [id] [int] NULL,
    [number] [varchar](50) NULL,
    [fruit] [varchar](50) NULL

INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (1, NULL, N'one')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (2, NULL, N'apple')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (3, NULL, N'banana')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (4, NULL, N'orange')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (5, NULL, N'two')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (6, NULL, N'apple')
INSERT [dbo].[Table1] ([id], [number], [fruit]) VALUES (7, NULL, N'banana')

+----+--------+--------+
| id | number | fruit  |
+----+--------+--------+
|  1 | NULL   | one    |
|  2 | NULL   | apple  |
|  3 | NULL   | banana |
|  4 | NULL   | orange |
|  5 | NULL   | two    |
|  6 | NULL   | apple  |
|  7 | NULL   | banana |
+----+--------+--------+

+----+--------+--------+
| id | number | fruit  |
+----+--------+--------+
|  1 | one    | one    |
|  2 | one    | apple  |
|  3 | one    | banana |
|  4 | one    | orange |
|  5 | two    | two    |
|  6 | two    | apple  |
|  7 | two    | banana |
+----+--------+--------+

基本上我想用水果列中的值更新数字中的空列，每当一个字符串出现不是'apple，banana，orange'时，我希望在输入新字符串之前保持输入该值水果。所以结果应该看起来像第二个例子。

我想我需要使用某种循环，但我不确定如何去做，到目前为止我的尝试是

declare @i varchar
set @i = 'one'

while @i = 'one' or @i not in ('apple', 'banana', 'orange')
begin
update Table1
set number = @i

set @i = fruit

end

但是在尝试设置@i = fruit

时出错

非常感谢任何帮助

Answer 1

declare @tb as TABLE (    [id] [int] NULL,    [number] [varchar](50) NULL,    [fruit] [varchar](50) NULL)

INSERT @tb ([id], [number], [fruit]) VALUES (1, NULL, N'one')
INSERT @tb ([id], [number], [fruit]) VALUES (2, NULL, N'apple')
INSERT @tb ([id], [number], [fruit]) VALUES (3, NULL, N'banana')
INSERT @tb ([id], [number], [fruit]) VALUES (4, NULL, N'orange')
INSERT @tb ([id], [number], [fruit]) VALUES (5, NULL, N'two')
INSERT @tb ([id], [number], [fruit]) VALUES (6, NULL, N'apple')
INSERT @tb ([id], [number], [fruit]) VALUES (7, NULL, N'banana')

select * from @tb

declare @count int = (select COUNT(*) from @tb)
declare @cursor int = 1
declare @updateValue as nvarchar(10)

While @cursor <= @count
BEGIN
set @updateValue = ISNULL((select fruit from @tb where id = @cursor and fruit not in ('apple', 'banana', 'orange')),@updateValue);
update @tb set number = @updateValue where id = @cursor
set @cursor = @cursor + 1;
END

select * from @tb

试试这个。

Answer 2

UPDATE dbo.Table1
SET number = t.fruit
FROM (SELECT t1.id, t2.fruit    
      FROM (SELECT id, 
                   MAX(CASE WHEN fruit NOT IN ('apple','banana','orange') 
                            THEN id ELSE 0 END)
                   OVER (ORDER BY id ROWS UNBOUNDED PRECEDING) AS i
           FROM dbo.Table1) t1
      INNER JOIN dbo.Table1 t2 ON t2.id = t1.i) t
WHERE t.id = dbo.Table1.id;

MAX表达式使用较新的（Sql Server 2012）windowing functions将id与最近的前一行的id配对，但不包含'apple'，'banana'或{ {1}}，然后它只是加入原始表并进行更新。

根据另一列中的值更新sqlserver中的列

2 个答案: