Question

I used a Java method called ‘containsAll()’ to check if ArrayLists have common content.

Let’s say I have a list A (one row), and several other lists in a MySQL table (in column ‘name’, row by row).
All lists consist of comma-separated Strings (at least one String in a list) – names or whatever.

Now, I want to check if all Strings in list A can be found in any of the rows in column ‘name’.
The result set should show all the rows in ‘name’ that match, that includes rows/lists must have all Strings in list A, and can have additional Strings.

Example I

A: ‘Mr.T’

____name_________________________________________
 'Hannibal'
 'Hannibal','Face','Murdock','Mr.T','Donald Duck'
 'Face','Donald Duck'
 'Superman','Chuck Norris','Mr.T'
_________________________________________________   

Result set: ‘Hannibal’,’Face’,’Murdock’,’Mr.T’,’Donald Duck’ -AND-
‘Superman’,Chuck Norris’,’Mr.T’

---

Example II

A: ‘Rocky’, ‘Mr.T’, ‘Apollo’

 ______name__________________________________________________
 'Hannibal','Face','Murdock','Donald Duck','Superman','Mr.T'
 'Rocky','Apollo','Ivan'
 'Apollo', 'Superman','Hannibal','Rocky','Mr.T','Chuck Norris'
 'Rocky','Mr.T','Apollo','Chuck Norris'
_____________________________________________________________   

Result set: ‘Apollo’, ‘Superman’,’Hannibal’,’Rocky’,’Mr.T’,’Chuck Norris’ -AND-
‘Rocky’,’Mr.T’,’Apollo’,’Cuck Norris’

I wonder if one can carry out those results using a MySQL query.
Thank you in advance.

Answer 1

It appears you want to do an array intersection, except your array is a single column. It can be done, but it will be slow, difficult to debug and will not leverage the power of relational databases. A better way would be to change your table schema to something like this:

Table groups

group_id int unsigned not null auto_increment primary key,
character_list text

Table members_in_group

group_id int unsigned not null,
group_member varchar(45) not null

Then you can query like this:

SELECT group_id, character_list
FROM groups g 
  JOIN members_in_groups m USING (group_id)
WHERE m.group_member IN ('Mr. T', ...);

The groups table is probably very like your current table. The members_in_groups table is the same data chopped up into easily searchable parts.

ETA given your comment, this should work if you can guarantee that each character_list contains only one instance of each character:

SELECT group_id, 
       SUM(CASE m.group_member IN ('Mr. T', 'Apollo', 'Rocky') THEN 1 ELSE 0 END) AS tally, 
       character_list
FROM groups g
  JOIN members_in_groups m ON (g.group_id=m.group_id)
GROUP BY group_id
  HAVING SUM(CASE m.group_member IN ('Mr. T', 'Apollo', 'Rocky') THEN 1 ELSE 0 END) = 3;

In this case the HAVING clause must equal 3 because there are 3 members in IN ('Mr. T', 'Apollo', 'Rocky').