Does anybody out there run a Google Custom Search Engine?

Still wondering how to save your user’s search queries?

If you host yourself your own CSE, and your server can run PHP/MySQL, here is the solution!

First of all, you have to set up a MySQL table where the search terms searched on your Google Custom Search Engines will be saved: to do so, open your favourite MySQL frontend (like PhpMyAdmin), and run this query:

CREATE TABLE `query` (
`id` int(11) NOT NULL auto_increment,
`query` varchar(255) NOT NULL default '',
`datetime` int(11) NOT NULL default '0',
PRIMARY KEY (`id`)
) ENGINE=MyISAM


Done? Well. Now, open the page in which you pasted the Google’s code that displays search results, and add to that page the code that will save data: it grabs search term from the URL:

< ?php
/*-----------------------
First part: db connection
-------------------------*/
$dbhost = "localhost";
$dbname = "mylittledb";
$dbuser = "root";
$dbpass = "password";
$db=mysql_connect($dbhost, $dbuser, $dbpass);
if ($db==FALSE)
die("Error while connecting to MYSQL ".mysql_error());
mysql_select_db($dbname ,$db);
/*------------------------------
Read and save the search query
-------------------------------*/
$querystat = mysql_real_escape_string($_GET['q']);
$datetime = time();
if( ($_SERVER['HTTP_REFERER'] == '') AND ( ($querystat != $_SESSION['prev_search']) OR ( ($datetime - $_SESSION['datetime']) > 60) ) ) {
$insertquery = "INSERT INTO `query` ( `query` , `datetime`) VALUES ( '$querystat' , '$datetime');";
mysql_query($insertquery, $db);
}
$_SESSION['datetime'] = $datetime;
$_SESSION['prev_search'] = $querystat;
?>

That’s all.

This code saves also the time/date of each search, so you can may later analyze searches e.g. per day, per month, or per time period.

Example’d code saves queries only if the referer is blank (i.e. the user did not request the search results page clicking on a link to it, but he has to manually submit the form), and it prevents multiple page refreshes (a minute at least has to pass, or the search query has to be different, before the term is saved into database).

If you don’t want filters, and you want to save right every query, keep just the database connection and the two lines inside the IF statement for the second part.

If you have installed MySQL server in the standard directory /usr/local/mysql, do not read this post. Else, you may want to know

how to use MySQL shared libraries on your Linux system

For organization purposes, I have installed MySQL in /home dir (which is a separate partition), precisely in /home/server/mysql. Thus, I keep mysql/var, directory (which contains databases) safe from heavy system changes.

My system did not recognized the shared library libmysqlclient, that is under /home/server/mysql/lib/mysql.
How do you tell Linux “Hey, you will find the library here, go and use it”?

Just edit the file /etc/ld.so.conf and add the path to your library. Then, reboot or give this command:

to register new libraries.

Compiling programs that require MySQL components

Another issue you’ll have to deal with when your MySQL is installed in a non-standard path, is the source compilation and use of programs that require some MySQL components.

For example, I use a C program that has this header on its source:

I got errors related to paths when compiling this source.
The solution is not changing the source in order to make #include lines match real paths, but consists in telling the compiler where mysql directory is. Compile with such a command line (or change your Makefile accordingly):