选择列在一行中包含三个大写字母的位置

Question

I'm using SQL Server 2019 and trying to debug my ProperCase function that converts strings to proper case.

I have a table which I used my function to create, containing all varchar columns, one of them is named Surname .

I want to select the rows where Surname contains three or more uppercase letters in a row.

I've searched this site and Google etc and there's plenty of examples for finding where there are any uppercase letters or no uppercase letters in a field, but this is a little more subtle than that.

The column can contain any total number of upper or lowercase letters, but I only want to select the rows where it contains three or more uppercase letters next to one another.

Could a regular expression work here?

Answer 1

Not really any regex support in SQL Server natively unless you want to install custom CLR objects. If your data is stored as case insensitive and you want to perform case sensitive searches, one way is to use the COLLATE clause against the column.

DECLARE @x TABLE(i int, surname nvarchar(500));

INSERT @x(i, surname) VALUES
    (1, 'this is not a match'),
    (2, 'this is a MATCH'),
    (3, 'this is not a match'),
    (4, 'this is DEFINITELY a match');
    
DECLARE @min int = 3;

SELECT i, surname
  FROM @x
  WHERE surname COLLATE Latin1_General_BIN2 
  LIKE N'%' + REPLICATE(N'[A-Z]', @min) + N'%';

Results:

i      surname
----   -----------------------------
2      this is a MATCH
4      this is DEFINITELY a match

This dbfiddle also demonstrates other values for @min (in case you want to identify 4, or 40, or 300 consecutive upper-case characters in a row).

Note this won't perform well, so hopefully it's not something you're doing often and at scale. Also it won't find a surname with other characters between the upper-case characters, like Van DE Moor or MC-Adams . Not that those are normal, but data isn't normal, and want you to understand the bill of goods.