Many desktop applications (e.g, those built with Delphi) use "Database aware components", such as a grid, which display the contents of a database - usually the result of a query - and auto-update their display when the database contents change.
Does such a thing exist for PHP (maybe displaying the result of a query in an HTML table, and updating)?
If not, how would we go about creating it?
(Note: This seemingly similar question didn't help me)
It's not technically possible to update a HTML page once rendered with pure PHP because of the static nature of the HTTP protocol so any solution would have to include JavaScript and AJAX calls.
Emulating using AJAX to re-render the table every 5 minutes
It wouldn't be hard to emulate though, just make a PHP page which gets the results of a database and puts them in a table, then use jQuery's .load() function to get that table and render it in a DIV of your choice on an interval of 5 seconds or whatever.
Something like:
function updateTable(){
$('#tableDiv').load(url);
}
var url = 'renderTable.php';
setInterval(updateTable,5000);
You can put this in any PHP (or HTML) page with a DIV with id tableDiv and it will render the output of renderTable.php in that div every 5 seconds without refreshing.
Actually monitoring the database
This is possible, you'd have to set up a PHP file on a cron for every 5 seconds (or an AJAX call every 5 seconds) to run something like SELECT MD5(CONCAT(rows,'_',modified)) AS checksum FROM information_schema.innodb_table_stats WHERE table_schema='db' AND table_name='some_table'; (assuming innoDB), you'd then compare this to the previous checksum.
If the checksums are identical, you could pass 'false' to your modified AJAX call, which would tell it not to render anything over the table. If they aren't, you could pass it the HTML of the new table to render in place.
it could be done but with mix of different technologies
if I would like to monitor in real time changes made in database I would think about triggers and sockets - trigger in database should call (on insert or update) function that will add event to queue - here's example function for postgresql (plsh is custom handler)
CREATE FUNCTION propagate_event(tablename) RETURNS text AS '
#!/bin/sh
# execute script that will add event to message queue
/var/scripts/change_in_table.sh "$1"
' LANGUAGE plsh;
client connects to socket and receives that events in real time
Related
What I would like to do is push JSON data from a PHP script to Vue (in the view of an MVC PHP framework), comparable to what you would see in an online exchange where the data updates in near real time. So without a page reload.
Now I understand you can use fetch to fetch data (and probably set an interval so it does that every x seconds), but I would like the UI to update in (near) real-time as the PHP script outputs the data. How to do this? Preferably without using a third party like pusher.com.
Some details about the UI:
It's about 10 to 30 rows in a <table>
In each row there is one status label and a few numbers that would get updated
There are a few other generic elements on the page that need an update (eg overall status label)
I have a database with variables I want to change depending on the date. The events in the database have dates. After this date I would like to set a variable 'past' to true. Meaning, the event has past. How can I approach this with php or is this something that needs to be added to the database itself (using phpMyAdmin) ?
Naturally, you first need code that does the actual update. This could very well be a PHP script, but you probably won't want to embed it into a web page, but to run it from the command line (see below).
Secondly, you need to execute that code on a regular basis so every update is done within a guaranteed time span after the event has passed. To execute scripts or other code regularly, you can use cron (if you are on Linux) and similar mechanisms.
If you need to make sure that the updates are done as fast as possible after the respective event has passed, you will need to write a full-blown daemon which runs in the background and does its checks 10 times per second, for example, or (instead of polling) uses appropriately programmed one-shot timers.
If you insist on writing the script in PHP and don't want to run it from the command line (as required by cron for example), you still can embed your code into a web page as usual and then use tools like wget or curl in conjunction with cron to retrieve that web page and get your code executed regularly.
Current situation: I have a web page that uses AJAX/JQUERY to refresh all the content on the page every 17 seconds. Every time this happens the server queries the database for data from two tables, one of which is large (450MiB in size, 11 columns) and processes all the data.
This is too resource intensive. I want:
The server queries the database only when one of the two tables have changed.
The page then reloads the page through AJAX only when the tables have been updated and the server has re-processed the data.
I think this falls under the category of comet programming. I'm not sure.
2 is easy. The webpage calls 'update.php' every 17 (or maybe less) seconds. The PHP script returns no data if no changes have been made. Only if data is returned then the current page is replaced with the new data. Please advise me if there is a better way.
As for 1 my googling tells that every time one of my two tables is updated I can put a notification in a table (or maybe just a single byte in a file) to indicate that I must query the database again and then the next time that the webpage sends an AJAX request I return the data.
The problem is that I'm working with a rather large code base I didn't write and I don't know of all the places that either of the two tables may be updated. Is there an easier way to check when the database is modified.
I'm working with PHP, Apache, Drupal and MYSQL.
You can chekout Server Sent Events
A server-sent event is when a web page automatically gets updates from a server.
there is an excellent article on HTML5rocks.com - Server Sent Events
All You have to do is create an object
var source = new EventSource('xyz.php'); //Your php files which will return the updates.
Once you create an object,you can listen to the events
source.addEventListener('message', function(e) {
console.log(e.data);
}, false);
Is there a way that I can query an Oracle 10g database, and display the results in a dynamically refreshed html file every 3 minutes, for example?
Here is my predicament: I have several queries that I would LOVE to display the results of to a whole organization on a basic HTML web page with some CSS. The problem is that I do NOT want a user to be able to constantly refresh a page in his/her browser, and thus severely bog down the database. I have no problem writing the queries, or writing the HTML and CSS needed to display the tables. It's almost as if I would like to query, export results to XML every 3 minutes, and constantly have an HTML or PHP file that is pointing to the dynamically updated XML file. I am open to other options as well...
I have basic user access with the Oracle DB...nothing Admin like. I do have access to a server, though, and have experience with PHP, PL/SQL, and HTML. Perhaps I would have to get into a lower level programming language like Python? I am kind of stuck here. Any kind of help would be appreciated!
you can also execute an Ajax Request every 3 minutes using the setTimeout() function.
Using jQuery framework
$(document).ready(function() {
setTimeout("getFeed()", 180000); //180000 = 3 minutes in milliseconds
});
function getFeed() {
//ajaxRequest here
}
For more info on ajax you can go here: http://api.jquery.com/jQuery.ajax/
Setup a materialized view(mv), point your app to this mv, and then setup a scheduler job to refresh it on whatever frequency you like.
See dbms_scheduler for setting up scheduler jobs in Oracle.
One note: you may want to do an atomic_refresh=>true to do deletes/inserts into the mv instead of truncate/insert (if atomic_refresh=>false, there will be 0 rows in mv until refresh is completed).
An simple example mv creation:
create materialized view MY_MV
tablespace MY_TS
build immediate
refresh complete on demand
with primary key
as
SELECT a.foo, b.bar
from table_a a, table_b b
where a.col1 = b.col2
and a.baz='BLAH'
;
An example refresh call:
dbms_mview.refresh('MY_MV', 'C', atomic_refresh=>true);
I was just wondering how the PHP is behaving in the background.
Say I have a PHP which creates an array and populates it with names.
$names = Array("Anna", "Jackson" .... "Owen");
Then I have a input field which will send the value on every keypress to the PHP, to check for names containing the value.
Will the array be created on every call? I also sort the array before looping through it, so the output will be alphabetical. Will this take up time in the AJAX call?
If the answer is yes, is there some way to go around that, so the array is ready to be looped through on every call?
There's no difference between an AJAX request and a "normal" http request. So yes, a new php instance will be created for each request. If the time it takes to parse the script(s) is a problem you can use something like APC.
If those arrays are created at runtime and the time this takes is a problem you might store and share the values between requests in something like memcache
No matter what method you use to create the array, if it's in the code, if you pull it out of a database, a text file or any other source, when the web server gets an http request, ( whether it be Ajax or not ) it will start the execution of the PHP script, create its space in memory, and the array will be created.
There's only one entry point for a PHP script and it's the first line of it, when an http rquest points to it. (or when another script is included, which is the same)
As far as I know then it will have to create the array each time as the AJAX will make a new server request on each key press on the input input field. Each server request will create the array if you create the script to do so.
A better method would be to use a database to store the names.
Yes it will be created and destroyed every time you run the PHP script.
If this is a problem you could look at persisting this data somewhere (e.g. in a Session or in a Database), but I would ask whether it is really causing you so much of a performance problem that you need to do this?
(it's not a direct answer to your question, but it can help, if you are concerned about performances)
You say this :
Then I have a input field which will
send the value on every keypress to
the PHP
In this case, it is common pratice to not send the request to the server as soon as a key is pressed : you generally wait a couple of milliseconds (between 100 ms and 150 ms, I'd say), to see if there is not another keypress in that interval :
if the user types several keys, he usually types faster than the time you are waiting, so, you only send a request for the last keypress, and not every keypress
if the user types 4 letters, you only do 1 request, instead of 4 ; which is great for the health of your server :-)
speaking of time for the user : waiting 100 ms plus the time to go to the server, have the script running, and get back from the server is almost the same as without waiting 100 ms first ; so, not bad for the user
As a sidenote : if your liste of data is not too big (20 names is definitly OK ; 100 names would probably be OK ; 1000 might be too much), you could store it directly as a Javascript array, and not do an Ajax request : it's the fastest way (no client-server call), and it won't load your server at all.