php mysql and in_array performance - php

Question

Is there anything to get more performance? Because it tooks about 8 secs. car_models database has over 100.000 records, cars database has over 20.000 records and finaly car_parts has over 20.000 records. I could do for that a database something like keywords. But the problem is the databases are dynamic. I mean the records adding everyday by editors. And I can't touch the part of the software because it is not open source...
$news = dbquery("SELECT * FROM " . DB_NEWS . " ORDER BY news_id DESC LIMIT 0,1");
while ($news_data = dbarray($news))
{
echo $news_data['news_subject'];
$news_first_part = str_replace("\\", "", $news_data['news_news']);
$news_first_part = explode('.', $news_first_part); //first phrase
$news_first_part = $news_first_part[0];
$news_second_part = str_replace("\\", "", $news_data['news_extended']);
$find = array();
$keywords = dbquery("SELECT name FROM cars GROUP BY name");
while ($keywords_data = dbarray($keywords))
{
$my_keyword = $keywords_data['name'];
$news_first_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_first_part);
$news_second_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_second_part);
if ($news_first_part OR $news_second_part AND !in_array($my_keyword, $find,true))
{
array_push($find, $my_keyword);
}
}
$my_keyword="";
$keywords = dbquery("SELECT name FROM car_models GROUP BY name");
while ($keywords_data = dbarray($keywords))
{
$my_keyword = $keywords_data['name'];
if (strlen($my_keyword) > 10 && !in_array($my_keyword, $find, true))
{
$news_first_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_first_part);
$news_second_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_second_part);
}
}
$keywords = dbquery("SELECT name FROM car_parts GROUP BY name");
while ($keywords_data = dbarray($keywords))
{
$my_keyword = $keywords_data['name'];
if (strlen($my_keyword) > 10)
{
$news_first_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_first_part);
$news_second_part = str_replace($my_keyword, "<a href='keyword.php?keyword=$my_keyword' title='$my_keyword'>$my_keyword</a>", $news_second_part);
}
}
$my_keyword="";
echo $news_first_part . '.'; //note I added the final ponctuation
$news_first_part .= ".";
$news_second_part = str_replace($news_first_part, "", $news_second_part);
echo nl2br($news_second_part);
}

Answer 1

You could build some kind of array tree-structure, to limit the size of the sub-arrays. The array would then become:
$arr['a']['apple'];
$arr['a']['anna'];
$arr['a']['awesome'];
$arr['b']['bread'];
$arr['b']['beer'];
$arr['c']['cucumber'];
Instead of
$arr['apple'];
$arr['anna'];
$arr['awesome'];
$arr['bread'];
$arr['beer'];
$arr['cucumber'];
As you can see, an in_array would become lots faster.
If we would like to check "php", there is no array for the first-letter "p": nothing has to be walked through again
If we would like to check "ananas", we would have to lookup 3 items, instead of 6!
Note: the "outer" array could consist of the 1st letter (or 2, 3, 4, etc.) depending on the size of your set.
In PHP code you would get something like
$word = "ananas";
$arr = array();
if (!isset($arr[$word{0}]) || !in_array($word, $arr[$word{0})) {
// New word
if (!isset($arr[$word{0}])) {
$arr[$word{0}] = array($word);
} else {
$arr[$word{0}][] = $word;
}
}

Answer 2

Try it this way, it is much faster.
if ($news_first_part OR $news_second_part AND !isset($find[$my_keyword]))
{
$find[$my_keyword] = 1;
}
Update the rest of the code correspondingly. Actually, you do not need to use a !isset($find[$my_keyword]) check here.