97.394570112228 Query OK, 1 row affected (43.05 sec) the overhead of parsing and network communication

mysql> create table w0904procedure (wa char, wb char, wd char, wi char);
Query OK, 0 rows affected (0.00 sec)

mysql> INSERT INTO w0904procedure VALUES(NULL, 0, 'www0904', 'w0904d');
ERROR 1406 (22001): Data too long for column 'wd' at row 1

mysql> desc w0904procedure;
+-------+---------+------+-----+---------+-------+
| Field | Type    | Null | Key | Default | Extra |
+-------+---------+------+-----+---------+-------+
| wa    | char(1) | YES  |     | NULL    |       |
| wb    | char(1) | YES  |     | NULL    |       |
| wd    | char(1) | YES  |     | NULL    |       |
| wi    | char(1) | YES  |     | NULL    |       |
+-------+---------+------+-----+---------+-------+
4 rows in set (0.00 sec)

mysql> DROP PROCEDURE IF EXISTS insert_many_rows_w0904;
Query OK, 0 rows affected (0.00 sec)

mysql> delimiter //
mysql> CREATE PROCEDURE insert_many_rows_w0904 (IN loops INT)
    -> BEGIN
    ->  DECLARE v1 INT;
    ->  SET v1=loops;
    ->  WHILE v1 > 0 DO
    ->          INSERT INTO w0904procedure VALUES(NULL, 0, 'w', 'w');
    ->          SET v1 = v1 - 1;
    ->  END WHILE;
    -> END;
    -> //
Query OK, 0 rows affected (0.00 sec)

mysql> delimiter ;
mysql>
mysql> SET @loops=987987;
Query OK, 0 rows affected (0.00 sec)

mysql> CALL insert_many_rows_w0904(@loops);
Query OK, 1 row affected (43.05 sec)



mysql> SELECT * FROM w0904procedure LIMIT 20;
+------+------+------+------+
| wa   | wb   | wd   | wi   |
+------+------+------+------+
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
| NULL | 0    | w    | w    |
+------+------+------+------+
20 rows in set (0.00 sec)
 

mysql> SELECT COUNT(*) FROM w0904procedure;
+----------+
| COUNT(*) |
+----------+
|   987988 |
+----------+
1 row in set (0.00 sec)

High Performance MySQL, Third Edition

by Baron Schwartz, Peter Zaitsev, and Vadim Tkachenko

 

<?php

$db = new mysqli('localhost', 'root', '', 'w');
if($db->connect_error){
    die('Connect Error ( '.$db->connect_errno.' )'.$db->connect_error);
}
echo 'Connect OK.<br>'; 
$start = microtime(TRUE);
if($res=$db->query('SELECT COUNT(*) FROM w0904procedure')){
    while($row=$res->fetch_all()){
        var_dump($row);
    }
}
$end0 = microtime(TRUE);
if($res=$db->query('SELECT * FROM w0904procedure LIMIT 2')){
    while($row=$res->fetch_all()){
        var_dump($row);
    }
}
$end1 = microtime(TRUE);
$w = 987987;
do{
    $db->query('INSERT INTO w0904procedure VALUES(NULL, 1, "w", "w")');
    $w--;
}
while($w>0);
$end2 = microtime(TRUE);
echo $end0 - $start, '<br>', $end1 - $end0, '<br>', $end2 - $end1;

Connect OK.

D:\wamp64\www\wdb.php:11:
array (size=1)
  0 => 
    array (size=1)
      0 => string '987988' (length=6)

D:\wamp64\www\wdb.php:17:
array (size=2)
  0 => 
    array (size=4)
      0 => null
      1 => string '0' (length=1)
      2 => string 'w' (length=1)
      3 => string 'w' (length=1)
  1 => 
    array (size=4)
      0 => null
      1 => string '0' (length=1)
      2 => string 'w' (length=1)
      3 => string 'w' (length=1)

0
0.003000020980835
97.394570112228

mysql> SELECT COUNT(*) FROM w0904procedure;
+----------+
| COUNT(*) |
+----------+
|   987988 |
+----------+
1 row in set (0.00 sec)

mysql> SELECT COUNT(*) FROM w0904procedure;
+----------+
| COUNT(*) |
+----------+
|  1975975 |
+----------+
1 row in set (0.00 sec)

mysql>

We usually prefer to keep stored routines small and simple. We like to perform complex logic outside the database in a procedural language, which is more expressive and versatile. It can also give you access to more computational resources and potentially

to different forms of caching.

However, stored procedures can be much faster for certain types of operations—especially when a single stored procedure call with a loop inside it can replace many small queries. If a query is small enough, the overhead of parsing and network communication becomes a significant fraction of the overall work required to execute it.

 

To illustrate this, we created a simple stored procedure that inserts a specified number

of rows into a table.

The stored procedure is much faster, mostly because it avoids the overhead of network communication, parsing, optimizing, and so on.