Question

I have a PHP script something like:

$i=0;
for(;$i<500;++i) {
   //Do some operation with files numbered 0 to 500;
}

The thing is, the script works and displays the end results, but the operation takes a while and watching a blank screen can be frustrating. I was thinking if there is some way I can continuously update the page at the client’s end, detailing which file is currently being worked upon. That is, can I display and continuously update what is the current value of $i?

The Solution

Thanks everyone! The output buffering is working as suggested. However, David has offered valuable insight and am considering that approach as well.

Answer 1

You can buffer and control the output from the PHP script.

However, you may want to consider the scalability of this design. In general, heavy processes shouldn’t be done online. Your particular case may be an edge in that the wait is acceptable, but consider something like this as an alternative for an improved user experience:

1. The user kicks off a process. This can be as simple as setting a flag on a record in the database or inserting some “to be processed” records into the data.
2. The user is immediately directed to a page indicating that the process has been queued.
3. An offline process (either kicked off by the PHP script on the server or scheduled to run regularly) checks the data and does the heavy processing.
4. In the meantime, the user can refresh the page (manually, by navigating elsewhere and coming back to check, or even use an AJAX polling mechanism to update the page) to check the status of the processing. In this case, it sounds like you’d have several hundred records in a database table queued for processing. As each one finishes, it can be flagged as done. The page can just check how many are left, which one is current, etc. from the data.
5. When the processing is completed, the page shows the result.

In general this is a better user experience because it doesn’t force the user to wait. The user can navigate around the site and check back on progress as desired. Additionally, this approach scales better. If your heavy processing is done directly on the page, what happens when you have many users or the data processing load increases? Will the page start to time out? Will users have to wait longer? By making the process happen outside of the scope of the website you can offload it to better hardware if needed, ensure that records are processed in serial/parallel as business rules demand (avoid race conditions), save processing for off-peak hours, etc.