SQL 仅在字符串中的关键字之后提取单个单词或数字

Question

SELECT 
 NOTE_ID    
,NOTE_DATE  
,NOTE_TEXT  
FROM NOTE.dbo.NOTE_TABLE

从上面的简单查询，我得到：

NOTE_ID	NOTE_DATE	NOTE_TEXT
54876	2020-10-12	Purpose: Traveling Salesman Visit Customer Name: John Doe Date Account
		Opened: 6/8/19 Customer Address: 428 S. Palm Tree Lane Lokey AZ 85546 Primary
		Account Rep: Hank Stank Customer Account #: 1234567 Customer Preferred
		Contact Time: Evenings Customer Owns/Rents: Owns Customer Employed: Yes
		Customer Military Discount: No Customer Returns: No

我希望能够仅提取客户帐户 #“1234567”（或者如果它为 null/“UNKNOWN” - 客户帐户 # 之后和客户首选联系时间之前的任何内容。我尝试了各种子字符串函数来获取客户帐户 #，但无法删除之后的所有内容。任何建议将不胜感激。

请注意，NOTE_TEXT 字段都在一行上。不知道为什么它显示在多行上。

Answer 1

对于复杂的字符串解析（SQL 确实不太擅长），我使用了 CROSS APPLY 的层，这样我一次可以找到一个。它确实帮助我跟踪每一层并逐步调试。

DROP TABLE IF EXISTS #Note
Create Table #Note (
    NOTE_ID INT PRIMARY KEY
    ,NOTE_DATE DATE
    ,NOTE_TEXT Varchar(1000)
)

INSERT INTO #Note
Values(54876,'2020-10-12', 'Purpose: Traveling Salesman Visit Customer Name: John Doe Date Account Opened: 6/8/19 Customer Address: 428 S. Palm Tree Lane Lokey AZ 85546 Primary Account Rep: Hank Stank Customer Account #: 1234567 Customer Preferred Contact Time: Evenings Customer Owns/Rents: Owns Customer Employed: Yes Customer Military Discount: No Customer Returns: No')
,(62345,'2020-11-11', 'Customer Account #: 123 blah') /*Value at beginning of text*/
,(67890,'2021-02-28', 'Blah Customer Account #: 456') /*Value at end of text*/
,(99999,'2022-01-01', 'No matching account #') 



DECLARE @SearchPhrase AS Varchar(20) = 'Customer Account #:';

SELECT *
    ,AccountNumberParsedFromNote = SUBSTRING(NOTE_TEXT,B.IDX1,C.IDX2-B.IDX1)
FROM #Note AS A
CROSS APPLY (SELECT IDX1 = NULLIF(CHARINDEX(@SearchPhrase, NOTE_TEXT),0) + LEN(@SearchPhrase) + 1) AS B /*Index of letter after search phrase*/
CROSS APPLY (SELECT IDX2 = CHARINDEX(' ',NOTE_TEXT + ' '/*Trailing space accounts for value at end of string*/,B.IDX1)) AS C /*Index of end of next word after search phrase*/ 

Answer 2

我确信可能有更好的方法来做到这一点，但它应该匹配短语Customer Account #:

DECLARE @searchPhrase AS CHAR(20) = 'Customer Account #: ';
SELECT IIF(PhraseExists > 0, SUBSTRING(CustomerNumNextPhrase, 0, CHARINDEX(' ', CustomerNumNextPhrase)), 'UNKNOWN') AS Account
FROM (SELECT CHARINDEX(@searchPhrase, NotesData) AS PhraseExists,
             REPLACE(NotesData, SUBSTRING(NotesData, 0, CHARINDEX(@searchPhrase, NotesData) + 20), '') AS CustomerNumNextPhrase
      FROM test.dbo.Notes) A;

如果找不到短语Customer Account #: ，则内部查询报告一个0，而在这种情况下，外部查询报告UNKNOWN。如果内部查询报告除 0 以外的任何内容，它会返回主字符串的子字符串，在该子字符串中它会切断 Customer Account #: 之前的所有内容以及 Customer Account #: 短语本身。然后外部查询获取下一个关键字的子字符串，直到下一个空格。