SQL like - select strings that contain unspecified characters

Question

First of all, apologies if this is a duplicate question. I've done my best to search but was unsuccessful, and I couldn't even properly word my question in terms of keywords!

One of my tables has a column Name (nvarchar) . I want to find out which rows contain special characters without explicitly listing those characters. "Special" characters in my case means anything not in:

a-z A-Z 0-9 _ @ . , ( ) % + -

So for example:

Row 1: 'asdf Asdf 0123'
Row 2: 'asdf (Asdf) 012/3'
Row 3: 'zxcv [234]'
Row 4: 'asdf #0123'

I want to select rows 2, 3 and 4.

The easiest way is to include the characters I don't want, for example square brackets and slash:

SELECT * FROM Table
WHERE Name LIKE '%[\]\[/]%' ESCAPE '\'

This returns rows 2 and 3, or if I use NOT LIKE, rows 1 and 4. However, I also want to find other characters which I may not have thought of (such as the #). Listing the characters that ARE wanted -

SELECT * FROM Table
WHERE Name NOT LIKE '%[a-zA-Z0-9_@. ,()%+-]%'

doesn't work either, as it returns 0 results since all rows contain at least 1 of those characters.

Is there a way to restrict the latter LIKE statement not to match any string that contains my desired characters, but rather strings that contain only the desired characters and nothing else?

Answer 1

In MySQL you can do:

WHERE Name REGEX '[^-a-zA-Z0-9.,()%+]'

Not sure if the same REGEX operator exists in SQL Server, but it probably has something similar.

Answer 2

SELECT * FROM tblWHERE PATINDEX('%[^a-zA-Z0-9]%',col) >1

Should return the rows that have non alpha numeric characters. You may need to alter the regex for your specific needs.

use tempdb
create table tbl ( col varchar(40) NULL)

insert into tbl VALUES ('lowercase')
insert into tbl VALUES ('UPPER')
insert into tbl VALUES ('0123')
insert into tbl VALUES ('special characters & ^')
insert into tbl VALUES ('special characters ! \/')
insert into tbl VALUES ('special characters _ + = ')
insert into tbl VALUES (NULL)

SELECT * FROM tbl WHERE PATINDEX('%[^a-zA-Z0-9]%',col) >1

drop table tbl