Database design for social network?

Using facebook as an example, each member have a set of 'friends'.

In the database, do you have a table with a field 'memberID' and another field for 'friendID'? This seems very brute-force and inefficient in terms of redundant data (i.e. if member has 10 friends, that's 10 entries).

---OR---

you could have an entry per member, and a field for friendIDs which could be a comma-separated string or something.. this seems more efficient in the database but would dramatically increase processing time to parse the string.

I would choose option 1).

Ideas? Opinions? Alternate options??

Thanks in advance.

Database design for social network?christian myspace

Option 1) is the way to go. You should add foreign key constraints to ensure consistency.

Database design for social network?wrestlers myspace myspace.com

You would have a "Members" table:

MemberID LastName FirstName

-------------- -------------- ---------------

10000000 Brown Kwame

10000007 Bryant Kobe

10000008 Jordan Michael

A friend is also a member.

Next, create a "Friends" table:

MemberID FriendID

-------------- ------------

10000000 10000007

10000000 10000008

The MemberID is the original member, and the FriendID column contains the MemberIDs of the Friends

Now check the Results

SELECT FriendID FROM Friends WHERE MemberID='1000000'

FriendID

-------------

10000007

10000008

This way, you normalize your tables. This way a member can have as many friends as possible and you don't have repeating fields in your tables.
That is going to be a tough decision, may be you can contact a database expert at website like http://oktutorial.com/ .