SQL - select where elements of a string match elements of another string - mysql

Question

I have a table field, let's call it pattern containing a list of comma separated values.
For example: '10,20,30,40,50'
I need to select * from said table, where at least one element from another similar string appears in that field too.
For example, say I have the string '10,50,70' A record whose pattern field is '10,20,30,50,70' should be selected, because 10 and 70 are present in '10,50,70'.`
Is there any way of doing this, except lots of OR where i check if pattern LIKE '%10%' OR pattern LIKE %50% OR pattern LIKE %70% ?

Answer 1

You can use FIND_IN_SET() as an alternative like below though as already suggested in comment, consider Normalizing your table.
SELECT * FROM table_name
WHERE FIND_IN_SET('10','10,20,30,50,70')
OR FIND_IN_SET('50','10,20,30,50,70')
OR FIND_IN_SET('70','10,20,30,50,70') ;

Answer 2

As mentioned by Patrick Hofman, normalising the database is the best solution.
If you really must stick with such a table layout then there are a couple of solutions.
You can split up the comma separated values and join the results based on that. Not that the cross joined tables are just generating a range of numbers (in this case 0 to 99), but this must equal or exceed the max number of delimited values. It will also be slow as no indexes can be used. Also large numbers of records are generated in memory to do this calculation.
Something like this:-
SELECT DISTINCT sub0.id, sub1.id
FROM
(
SELECT DISTINCT id, SUBSTRING_INDEX(SUBSTRING_INDEX(pattern, ',', units.i + tens.i * 10), ',' -1) AS pattern_id
FROM some_table
CROSS JOIN (SELECT 1 AS i UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9 UNION SELECT 0) units
CROSS JOIN (SELECT 1 AS i UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9 UNION SELECT 0) tens
) sub0
INNER JOIN
(
SELECT DISTINCT id, SUBSTRING_INDEX(SUBSTRING_INDEX(pattern, ',', units.i + tens.i * 10), ',' -1) AS pattern_id
FROM some_table
CROSS JOIN (SELECT 1 AS i UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9 UNION SELECT 0) units
CROSS JOIN (SELECT 1 AS i UNION SELECT 2 UNION SELECT 3 UNION SELECT 4 UNION SELECT 5 UNION SELECT 6 UNION SELECT 7 UNION SELECT 8 UNION SELECT 9 UNION SELECT 0) tens
) sub1
ON sub0.pattern_id = sub1.pattern_id
AND sub0.id != sub1.id
If the comma separated value is the key to another table then you could maybe do a couple of joins based on FIND_IN_SET.
SELECT DISTINCT a.*, b.*
FROM some_table a
INNER JOIN some_patterns b
ON FIND_IN_SET(b.id, a.patterns)
INNER JOIN some_table c
ON FIND_IN_SET(b.id, c.patterns)
In reality it would probably be best to normalise you database, possibly using SQL based on these solutions just to help you extract the data to your new normalised format

Answer 3

I've inherited this method of storing small lists of integers and wrote the included function to rotate CSV of integers into a column.
CREATE function [dbo].[udf_IntegerColumnFromCSV] ( #DelimStr nvarchar(max) )
returns #ListOfInts table ( Value int )
AS
BEGIN
set #DelimStr = #DelimStr + ',' -- add one more to the end
declare #i int -- the "start" of the next value to be read
declare #j int -- the location of the next comma
declare #le int -- line end
set #i = 1
set #j = 1
set #le = len(#DelimStr)
while ( #j < #le ) begin
set #j = charindex( ',', #DelimStr, #i ) -- find the end of the next row
insert into #ListOfInts (Value) values ( cast(substring( #DelimStr, #i, #j-#i ) as integer) )
set #i = #j + 1
end
return
end
In a situation where you have a list of values stored as CSV:
select ID, ValueList from TableX
ID : ValueList
20 : 38, 39, 40, 41, 42, 44, 61,62,63,64,65,66, 70,71,72
21 : 12
22 : 61,62,63,64,65,66
Use the function to pivot the values into a column:
select ID, vl.* from TableX cross apply dbo.udf_IntegerColumnFromCSV(ValueList) vl
20: 38
20: 39
20: 40
20: 41
20: 42
20: 44
20: 61
20: 62
20: 63
20: 64
20: 65
20: 66
20: 70
20: 71
20: 72
21: 12
22: 61
22: 62
22: 63
22: 64
22: 65
22: 66