高级SQL查询（查询和比较中的查询）

Question

:)

I'm right now scratching my head while I try to figure out one or two complex MySQL queries I'm in need of.

Table "timestamps" consists of columns: name, login_time, logout_time

An explanation of what I'm trying to achieve is this:

in short: search for name X. Find all that NEVER EVER have been online at the same time ever. If so, if login_time is equal to the comparing posts logout_time OR the comparing posts login_time is equal to the searched name's logout_time, it gains a "point" for probability.

The query at first should fetch all rows with name X that the logout_time is NOT NULL.

Then using those results (will be multiple rows), compare all results to ALL other rows in the table where the logouttime is NOT NULL and compare the login_time and logout_time to never EVER been online at the same time.

With that result, (those that haven't been online at the same time EVER against the name first searched) finally check if anyone has the first searched name login_time is equal to the comparing posts logout_time OR the comparing posts login_time is equal to the searched name's logout_time.

Right now I'm using many, many single queries in Java against the MySQL server doing all the checks. Here is the method for comparing online times I've used and confirmed working.

p1In, P1Out is the searched name's login_time and logout_time p2In, P2Out is login_time and logout_time of the one it's currently comparing against (one that have logout_time set(NOT NULL))

        public boolean OnlineAtTheSameTime(long p1In, long p1Out, long p2In, long p2Out) {
    boolean result = false;
    result = ((p1In > p2In || p1Out < p2Out) && (p1Out < p2In || p1Out > p2Out) && (p1In < p2In || p1In > p2Out));
    return result;
}

after that I later just check

if (p1In == p2Out || p2In == p1Out)

//give p2 a "point" for probability

I will later use this information, using probability, find out who's who thanks to them logging in/out before/after each other while at the same time never ever been online at once.

Due to the "massive" amounts of unnecessary queries I'm making, I'm wondering if all this is possible with one or more queries that probably will save me lots of database load and sheer amount of queries.

Thanks a lot and looking forward to see what you can come up with. (if I could I would upload an image of a row and the table structure)

Answer 1

1. Find the ids of everyone who has EVER been online at the same time.

2. Use that as a subquery in the following to find the NEVER EVERs:

    FROM real_table LEFT JOIN ( SELECT id ... ) AS ever USING (id) WHERE ever.id IS NULL