protecting against data in db being corrupted by different charsets

schwim

Hi there guys,

I have someone using one of my scripts in a non-english environment. The script allows submissions of testimonials, then displays them in a random block.

He states that when he submits the data into the form, it shows up broken in the admin pane and in the random block.

I was going to suggest that he change the charset of the db in phpMyAdmin, but he states that he does not have access to a db management system.

My questions:

1) between a charset declaration in the page header and setting the charset of the database, am I completely covered against data corruption, or is there something else I need to do?

2) Can I alter the charset of db tables without having access to a db manager?

thanks,
json

schwim

After reading, I guess that I'm calling it by the wrong name when dealing with the database. What I need to alter I guess, is the collation.

If anyone has some thoughts on modifying existing tables to alter collation, as well as whether my steps to protect against data corruption are enough, I would greatly appreciate the help.

thanks,
json

schwim

Hey, I figured out how to alter the collation by looking at phpMyAdmin commands, so I figured that part out.

All I need help with now is #1. I need to find out how to protect against data corruption.

thanks,
json

MarkR

If your application is correctly build, data corruption will not occur.

You need to make sure:

All data going to and from the user is in the right charset. If you're not sure what to use, use utf8.
All data internally to PHP is in the right charset. Ideally use the same one (hint: utf8)
Data going to/from mysql is in the right charset. To avoid unwanted conversion, just use the same one
Data stored in the DB is in the right charset. Just use the same one (hint: utf8)
is achieved with appropriate HTTP and email headers etc
Means when you read/write files on the disc (including PHP files), they should be in the right charset already.
Means you should do SET NAMES utf8 on every connection (for maximum safety)
Means all your tables should be in utf8. I mean all of them, not just a few. To migrate, use ALTER TABLE (I really mean it).

Don't use phpmyadmin for anything, it's a steaming pile and will corrupt data.

If you have data corruption, I blame phpmyadmin, which is a dangerous but appealing toy and should NOT be used on production databases under any circumstances.

Mark

schwim

Hi there Mark,

the data corruption in this case was not phpMyAdmin's fault, simply because as I stated in the initial post, it was not used.

So, will

1) setting the ISO charset in the lang file
2) ensuring that the collation is correct in the db using ALTER

ensure against corrupted data, providing I've not written my code incorrectly elsewhere?

thanks,
json