Question

查询以下两个表：

CREATE TABLE [dbo].[MTCorrelations](



[CorrelationID] [int] IDENTITY(1,1) NOT NULL,
[StockA] [nvarchar](5) NOT NULL,
[StockB] [nvarchar](5) NOT NULL,
[Correlation] [float] NOT NULL,
[LengthStr] [nvarchar](5) NOT NULL,
[Date] [datetime] NOT NULL


)

CREATE TABLE [dbo].[Industries](

[IndustryID] [int] IDENTITY(1,1) NOT NULL,
[Symbol] [nvarchar](5) NOT NULL,
[Sector] [nvarchar](50) NULL,
[Industry] [nvarchar](50) NULL


)

with this query:

[CorrelationID] [int] IDENTITY(1,1) NOT NULL, [StockA] [nvarchar](5) NOT NULL, [StockB] [nvarchar](5) NOT NULL, [Correlation] [float] NOT NULL, [LengthStr] [nvarchar](5) NOT NULL, [Date] [datetime] NOT NULL

结果产生重复，因为该表有重复项，其中StockA与StockB相关性列在一行中，而相同的相关性列在另一行中，StockB在StockA列中，反之亦然。

由于每个相关性都列出了两次，我想添加一个where子句来将结果限制为stockA在stockB之前按字母顺序排列的结果。我试过了一个＆lt;在stockA和stockB之间它并没有起作用。 SQL是否有字符串的比较运算符？

Answer 1

为什么不这样做，它更有效率，不会产生A / B的排他性：

SELECT StockA, StockB, Correlation, LengthStr From MTCorrelations
WHERE StockA < StockB AND -- This is to remove the permutations
  EXISTS                  -- Fast check for StockA being within constraints
  (SELECT * 
   FROM Industries 
   WHERE Industry = 'Money Center Banks' AND 
         Symbol = StockA) AND
  EXISTS                  -- Fast check for StockB being within constraints
  (SELECT * 
   FROM Industries 
   WHERE Industry = 'Money Center Banks' AND 
         Symbol = StockB)      
ORDER BY Correlation DESC

如何按字母顺序比较SQL中的字符串？

1 个答案: