I have two tables tbl_user and tbl_projects
tbl_user
+-----+------+--------+
| id | name | skills |
+-----+------+--------+
| u1 | x | s1,s2 |
| u2 | y | s2,s3 |
| u3 | z | s3,s1 |
+-----+------+--------+
tbl_projects
+-----+--------+
| id | kills |
+-----+--------+
| p1 | s2 |
| p2 | s1,s3 |
| p3 | s3 |
+-----+--------+
For my application I want a sql query for list out all projects those which are match with the user skill
example, If i select the user u1 the result will be like
+-----+--------+
| id | kills |
+-----+--------+
| p1 | s2 |
| p2 | s1,s3 |
+-----+--------+
People better than I can help you build an SQL query to get you there. The way your tables look though, you will always have to build complex queries to pull out useful information. My advice, especially if you are still early in your project, is to change your table structure to make querying easier.
For instance, the tables below reflect the info in your OP in a structure that will make your life much much easier.
tbl_users: holds details that have a one-to-one relationship with each user
userID|name|email...
u1 | x | ...
u2 | y | ...
u3 | z | ...
tbl_skills: details that have a one-to-one relationship with each skill
skillID
s1
s2
s3
tbl_projects: details that have a one-to-one relationship with each project
pID| title | deadline
p1 | project a | 2016-08-15
p2 | project b | 2017-01-01
p3 | project c | 2015-08-22
tbl_user_skills: Each record has one user and one skill, both of which are foreign keys to this table (primary keys in tbl_users and tbl_skills respectively). It should have a UNIQUE index on (userID,skillID) to prevent duplicate entries.
userID|skillID
u1 | s1
u1 | s2
u2 | s2
u2 | s3
u3 | s1
u3 | s3
tbl_project_skills Each record has one project and one skill, both of which are foreign keys to this table (primary keys in tbl_project and tbl_skills respectively). It should have a UNIQUE index on (pID,skillID) to prevent duplicate entries.
pID|skillID
p1 |s2
p2 |s1
p2 |s3
p3 |s3
Once everything is organized this way, your queries will be much faster and much simpler to build. In fact, if you understand bitflag operations, you could condense this considerably (eg: have all skills from a user as one field in tbl_users but instead of s1,s2 you would use bits).
To get all projects with the skills of user u1:
SELECT p.pID, p.title
FROM tbl_projects p
LEFT JOIN tbl_project_skills ps ON p.pID = ps.pID
LEFT JOIN tbl_user_skills us ON ps.skillID = us.skillID
WHERE us.userID='u1'
GROUP BY p.pID
Result
pID| title
p1 | project a
p2 | project b