Query SQL Server with wildcard IN the database - sql

Question

I need to build a query where the criteria must match with wildcard in the database.
With an example it will be clearest.
I have a column with a field like this 963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX.
The ~ char is a wildcard.
So the following criterias must match :
~63-4-AKS~M
963-4-AKS1M
963-4-AKS~M2RN21AXA150AAA
963-4-AKSAM2RN21AXA150AAA
963-4-AKSCM2RN21AXA150A060C1D1DSDXX
963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX
I've tried so much things my head hurt :(
In the other way (with wildcard from the criteria) no problem, easy. But in this way I cannot find the key.
The problem is when I have a ~ in the field it doesn't match. So here only the first and last match with the following statement :
SELECT myField FROM myTable WHERE myField LIKE REPLACE('%' + myCriteria + '%', '~', '_');

Answer 1

It seems the patterns and the field are adjusted to the left.
If this is indeed the case, with my head bowed (full of sadness), here is a function.
create function is_a_match (#myField varchar(100),#myCriteria varchar(100))
returns bit
as
begin
declare #i int = 0
,#is_a_match bit = 1
,#len_myField int = len(#myField)
,#len_myCriteria int = len(#myCriteria)
,#myField_c char(1)
,#myCriteria_c char(1)
While 1=1
begin
set #i += 1
if #i > #len_myCriteria break
if #i > #len_myField
begin
set #is_a_match = 0
break
end
set #myField_c = substring(#myField ,#i,1)
set #myCriteria_c = substring(#myCriteria,#i,1)
if not (#myField_c = '~' or #myCriteria_c = '~' or #myField_c = #myCriteria_c)
begin
set #is_a_match = 0
break
end
end
return #is_a_match
end
GO
select myCriteria
,dbo.is_a_match (myField,myCriteria) as is_a_match
from (values ('~63-4-AKS~M' )
,('963-4-AKS1M' )
,('963-4-AKS~M2RN21AXA150AAA' )
,('963-4-AKSAM2RN21AXA150AAA' )
,('963-4-AKSCM2RN21AXA150A060C1D1DSDXX' )
,('963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX' )
,('963-4-AKS~M2RN21AXA150~~~0C1X1D~~XX' )
,('963-4-AKS~M2RN21AXA150~~~0C1D1D~~XXYY')
) c (myCriteria)
,(values ('963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX' )
) f (myField)
+---------------------------------------+------------+
| myCriteria | is_a_match |
+---------------------------------------+------------+
| ~63-4-AKS~M | 1 |
+---------------------------------------+------------+
| 963-4-AKS1M | 1 |
+---------------------------------------+------------+
| 963-4-AKS~M2RN21AXA150AAA | 1 |
+---------------------------------------+------------+
| 963-4-AKSAM2RN21AXA150AAA | 1 |
+---------------------------------------+------------+
| 963-4-AKSCM2RN21AXA150A060C1D1DSDXX | 1 |
+---------------------------------------+------------+
| 963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX | 1 |
+---------------------------------------+------------+
| 963-4-AKS~M2RN21AXA150~~~0C1X1D~~XX | 0 |
+---------------------------------------+------------+
| 963-4-AKS~M2RN21AXA150~~~0C1D1D~~XXYY | 0 |
+---------------------------------------+------------+
You are mixing between the field and the patterns.
The field may not hold wildcards.
E.g.
This is not a match because of the 'A's
963-4-AKS~M2RN21AXA150~~~0C1D1D~~XX
963-4-AKSAM2RN21AXA150AAA

Answer 2

If you could tighten the constraints on the wildcard, you might have a fighting chance here. What I mean is that you generate the valid permutations if a wildcard is presented in the persisted data. Then query the permutations with your existing query.
But if every wildcard has 36 possible options, this becomes exponentially painful.