nettika/phpmyadmin
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[import] Speed optimization to be able to import the sakila database

Browse Source
This commit is contained in:
Marc Delisle
2008-07-14 15:28:58 +00:00
parent d637fab135
commit 29d7c3e2f8
2 changed files with 18 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
ChangeLog
View File

@@ -66,6 +66,7 @@ danbarry
+ [GUI] Color picker (work in progress, now just for original theme)
- bug #1970836 [parser] SQL parser is slow, thanks to Christian Schmidt
+ rfe #1692928 [transformation] Option to disable browser transformations
+ [import] Speed optimization to be able to import the sakila database
2.11.8.0 (not yet released)
- patch #1987593 [interface] Table list pagination in navi,

64
libraries/import/sql.php
View File

@@ -48,6 +48,8 @@ $sql = '';
$start_pos = 0;
$i = 0;
$len= 0;
$big_value = 2147483647;
if (isset($_POST['sql_delimiter'])) {
$sql_delimiter = $_POST['sql_delimiter'];
} else {
@@ -86,72 +88,40 @@ while (!($GLOBALS['finished'] && $i >= $len) && !$error && !$timeout_passed) {
}
// Current length of our buffer
$len = strlen($buffer);
// prepare an uppercase copy of buffer for Windows because at least
// on PHP 5.2.5, stripos() is very slow
if (PMA_IS_WINDOWS) {
$buffer_upper = strtoupper($buffer);
}
// Grab some SQL queries out of it
while ($i < $len) {
$found_delimiter = false;
// Find first interesting character, several strpos seem to be faster than simple loop in php:
//while (($i < $len) && (strpos('\'";#-/', $buffer[$i]) === FALSE)) $i++;
//if ($i == $len) break;
$oi = $i;
$big_value = 2147483647;
$first_quote = strpos($buffer, '\'', $i);
if ($first_quote === FALSE) {
$first_quote = $big_value;
}
$p2 = strpos($buffer, '"', $i);
if ($p2 === FALSE) {
$p2 = $big_value;
// Find first interesting character
$old_i = $i;
// this is about 7 times faster that looking for each sequence i
// one by one with strpos()
if (preg_match('/(\'|"|#|-- |\/\*|`|(?i)DELIMITER)/', $buffer, $matches, PREG_OFFSET_CAPTURE, $i)) {
// in $matches, index 0 contains the match for the complete
// expression but we don't use it
$first_position = $matches[1][1];
} else {
$first_position = $big_value;
}
/**
* @todo it's a shortcoming to look for a delimiter that might be
* @todo we should not look for a delimiter that might be
* inside quotes (or even double-quotes)
*/
// the cost of doing this one with preg_match() would be too high
$first_sql_delimiter = strpos($buffer, $sql_delimiter, $i);
if ($first_sql_delimiter === FALSE) {
$first_sql_delimiter = $big_value;
} else {
$found_delimiter = true;
}
$p4 = strpos($buffer, '#', $i);
if ($p4 === FALSE) {
$p4 = $big_value;
}
$p5 = strpos($buffer, '--', $i);
if ($p5 === FALSE || $p5 >= ($len - 2) || $buffer[$p5 + 2] > ' ') {
$p5 = $big_value;
}
$p6 = strpos($buffer, '/*', $i);
if ($p6 === FALSE) {
$p6 = $big_value;
}
$p7 = strpos($buffer, '`', $i);
if ($p7 === FALSE) {
$p7 = $big_value;
}
// catch also "delimiter"
// stripos() very slow on Windows (at least on PHP 5.2.5)
if (! PMA_IS_WINDOWS) {
$p8 = stripos($buffer, 'DELIMITER', $i);
} else {
$p8 = strpos($buffer_upper, 'DELIMITER', $i);
}
if ($p8 === FALSE || $p8 >= ($len - 11) || $buffer[$p8 + 9] > ' ') {
$p8 = $big_value;
}
// set $i to the position of the first quote, comment.start or delimiter found
$i = min($first_quote, $p2, $first_sql_delimiter, $p4, $p5, $p6, $p7, $p8);
unset($first_quote, $p2, $p4, $p5, $p6, $p7, $p8);
$i = min($first_position, $first_sql_delimiter);
if ($i == $big_value) {
// none of the above was found in the string
$i = $oi;
$i = $old_i;
if (!$GLOBALS['finished']) {
break;
}