CODEWITHSUNDEEP
phpmysql
zookastos
zookastos

Reputation: 935

MySQL to search using joins

I have 2 mysql tables -

candskill - (cis, sid) - where cid = candidate id, and sid = skill id

Data in candskill (size - 257,000) -

c1, s1
c1, s2
c2, s3
c1, s4
c2, s5
...

skills - (sid, name) - where sid = skill id, and name = skill name

Data in skills (size 257,000)-

s1 - oracle
s2 - project management
s3 - oracle
s4 - testing
s5 - testing
...

Now, I want to fetch all the candidates who have skills 'oracle' and 'testing' both. Or I want candidate who have skills either 'oracle' or 'testing'. I want to have any AND/OR combination of skills present, and want to fetch candidates for those skills.

How would I achieve that?

This is what I have so far, which is not working in all scenarios.

select distinct(cs.cid), s.name from candskill cs 
inner join skills s on (cs.sid = s.sid and (s.name = 'oracle' or s.name = 'testing'))

Also, the query execution is taking too much time. approx 120 sec. How do we go about doing that.

I am thinking of writing a query, and passing the skill part of the query via php code, concate the strings, and generate new query each time a user searches for candidates for a particular skill.

Upvotes: 0

Views: 63

Answers (2)

Olaf Dietsche
Olaf Dietsche

Reputation: 74018

Maybe reducing the set of skills helps with the performance, e.g.

select cs.cid
from (select sid from skills where name in ('oracle', 'testing')) s
join candskills cs on cs.sid = s.sid

Instead of joining 250,000 x 250,000 rows, this will join 2 x 250,000 rows.

Furthermore, adding an index on skills.name and another on skills.sid and candskills.sid might improve the query further.

Upvotes: 1

ScaisEdge
ScaisEdge

Reputation: 133360

You could use an having clause on count of s.name 

  select cs.cid
  from candskill cs 
  inner join skills s on (cs.sid = s.sid and  s.name in (  'oracle' , 'testing'))
  group by cs.cid
  having count(distinct(s.name)) = 2

for 1 or 2 

  select cs.cid
  from candskill cs 
  inner join skills s on (cs.sid = s.sid and  s.name in (  'oracle' , 'testing'))
  group by cs.cid
  having count(distinct(s.name)) >= 1

Upvotes: 1

Related Questions