you are to process it by means of intval: $id = intval($id)).

It will be shorter: $id = (int) $id. And you also have forgotten about a rather useful thing - LIMIT

$somefield = str_replace("'", "'", $somefield).

Hmmm... Maybe it’s right to write $somefield = str_replace("'", "", $somefield)?

What for do you need functions if this navigation string is used on the page once only? Isn’t it simpler to include or request pages.inc which draws the necessary string and get current page along with number of pages within a request in the main page? It will also enable you to take into account specificity of request – filtration, sorting and it is also useful to indicate ‘page 5 of 2b’ at the top of a page.
You are not going to verify number of pages – such a client can always be found who is interested in entering a big number with hands.

Navigation string can be used twice – if there are two lists on a page (although it is a rather sophisticated variant    ) The thing is that include and require for data output are forbidden by me as far as I use templates. Following is done there:

$page->assign(array("limit" => get_limit(...), "nav_bar" => draw_bar(...)));

Last edited by Stre@m (April 10, 2007 7:03 am)

if ($count/2==intval($count/2))

Replace it with $count & 1. After this I’m not likely to read the rest.   

$amount = @mysql_result(mysql_query("SELECT count(id) as goods_total FROM goods"),0);

What is this?
list ($amount) = mysql_fetch_row (...); Isn’t it better?

Last edited by Stre@m (April 10, 2007 7:32 am)

It’s a perfect script. Thanks! The only bug has been mentioned when working (or isn’t it a bug?) – if the general lines’ number is less than number of lines on the page, all lines from the table are put out…

I have such a question. What is more productive in this case: to use LIMIT and get 2 queries to the base as a result (one of them is count and the other one is basic) or select all entries at once and do mysql_data_seek afterwards?

I have lots of news on my site. I divide them among pages (each page having N news). News are sorted on date in drawback order (the newer ones come first). In some time the very first news is going to appear on M page which means that URL will look like page.php?page_number=M Some time ago this news was on M-K page and it was exactly the moment when it was indexed by search engines. Then it is going to display this URL that hasn’t been possessing this news for a long time. Thus someone is very unlikely to be satisfied with such pages division method. Somehow we are to make URL look like page.php?page_where_first_news_id=L. Then a query will look like select (...) from news where news_id < $page_where_first_news limit N. It’s clear how to get URLs of the previous and the next page. But I cannot realize how to calculate URL of the page I. Are there any ideas about how to manage it with the minimal queries’ number?

Base is a directory in which files are paced (three files for each table). All of them have got extensions.

And reducing of loading on PHP due to simplifying of syntactic constructions.

cplusplus, look up documentation. We can see:

I’ve got such a problem… Paginal sorting has already passed into history. I’d like to create news’ archive on years and months. News’ date is of the DATE type in format YYYY-MM-DD... How can I realize it? Or should I make every page separately?

This all is good only in case when there are few entries in the base. And what if I have 100000 entries? Can you imagine that each string will contain pages’ numbers? It will occupy at least half the screen. But you are not to suggest something like displaying greater number of strings on a page. There are some interesting solutions but I’d like something better and nicer. Who would like to write a universal script for displaying pages?

if ($count/2==intval($count/2))

It's not the best variant. It isn't nice!!!
I do it like this:

$bgcolor=array("#eeeeee", "#fafafa");
for ($i=0, $y=count($post); $i<$y; $i++) { 
print "<tr bgcolor=".$bgcolor[($i % 2)]."><td></td></tr>"; 
}//for

It's not a problem to apply this idea here...
The code of my own works without any problems. At this it is 5 times shorter and it doesn't have bugs which were discussed above...