Question

Is the syntax of this query

SELECT * FROM  table1
WHERE var_c IN(
SELECT var_a FROM table2 
WHERE var_b =55554444
);

Equivalent to this?

SELECT table1.* FROM table1, table2
WHERE (table2.var_a=table1.var_c AND table2.var_b=55554444);

The first one takes about 7-8 seconds to run and the 2nd one takes about .75 seconds to run. When I use a Join statement it takes about 4-5 seconds to run.

Also is the syntax of this

DELETE FROM  table1
WHERE var_c IN(
SELECT var_a FROM table2 
WHERE var_b =55554444
);

And this:

DELETE table1.* FROM table1, table2
WHERE (table2.var_a=table1.var_c AND table2.var_b=55554444);

The same?

Answer 1

No, the 2 queries are not equivalent. The second can return duplicate rows, if (var_a, var_b) is not UNIQUE in table2

The first query though:

SELECT * 
FROM  table1
WHERE var_c IN
      ( SELECT var_a 
        FROM table2 
        WHERE var_b =55554444
      )

and if table2.var_a does not contain any NULL values, then it’s equivalent to this:

SELECT table1.* 
FROM table1
   , table2
WHERE table2.var_a = table1.var_c 
  AND table2.var_b = 55554444
GROUP BY table1.PK                --- Primary Key of table1

(which is better to be written with explicit JOIN syntax as:

SELECT table1.* 
FROM table1
  JOIN table2
    ON table2.var_a = table1.var_c 
WHERE table2.var_b = 55554444
GROUP BY table1.PK                --- Primary Key of table1

and this:

SELECT * 
FROM  table1
WHERE EXISTS 
      ( SELECT *
        FROM table2 
        WHERE table2.var_b = 55554444
          AND table2.var_a = table1.var_c
      )

---

About performance, the best advice you could get is to test, test and test again with your data. Try with different indexes and figure how to read the EXPLAIN plans.

You’ll probably find that IN (SELECT ... FROM ...) has not the best performance with current MySQL optimizer (although I hear that MariaDB plans some major improvements in next release, Maria 5.3) and that JOIN and EXISTS variants perform usually better.

But this heavily depends on the indexes you have on the tables. Without any index, all of them will be slow. And query time < 1 sec does not mean it’s fast. With a billion rows in the tables, yes, it’s pretty good. For a table with only a few thousand rows, time < 0.01 sec should be expected.