Reputation: 1328
I am doing a login script in php and html. The username and password dont need to be encrypted so I was trying to store them directly into the database (MySQL).
But some of my users have username and password containing special characters like "é" and must be case sensitive.
I have this table:
USER username: "ÉleCTeur", password: "ÉleCTeur", username: "Paul", password: "Paul"
I tried this query that work fine with "Paul" but return nothing with "ÉleCTeur".
SELECT
username,
password
FROM
user
WHERE
BINARY username = 'ÉleCTeur'
AND BINARY password = 'ÉleCTeur'
Did I just miss something or do BINARY
does not compare special characters?
What methode could I use to resolve this issue?
Upvotes: 2
Views: 1216
Reputation: 51980
You probably have inconsistent encoding between your db and the client program?
I take some time testing from MySQL CLI:
mysql> create TABLE user (username CHAR(20), password CHAR(20)) DEFAULT CHARSET=latin1 DEFAULT COLLATE=latin1_general_ci;
Query OK, 0 rows affected (0.25 sec)
mysql> insert into user value("ÉleCTeur", "ÉleCTeur");
Query OK, 1 row affected (0.00 sec)
mysql> select * from user;
+----------+----------+
| username | password |
+----------+----------+
| ÉleCTeur | ÉleCTeur |
+----------+----------+
1 row in set (0.00 sec)
mysql> select count(*) from user where username = "EleCTeur" and password = "EleCteur";
+----------+
| count(*) |
+----------+
| 0 |
+----------+
1 row in set (0.00 sec)
mysql> select count(*) from user where username = "ÉleCTeur" and password = "ÉleCteur";
+----------+
| count(*) |
+----------+
| 1 |
+----------+
1 row in set (0.00 sec)
EDIT: OK, this was a (too) quick and (far too) dirty test. As Sebastien noticed it, I use case insensitive collation here. So:
mysql> select count() from user where username = "éleCTeur" and password = "éleCteur"; +----------+ | count() | +----------+ | 1 | +----------+ 1 row in set (0.01 sec)
The real answer here would be to use an case sensitive (_cs
) collation. To know which one are available:
mysql> show collation where collation like '%_cs';
+--------------------+---------+----+---------+----------+---------+
| Collation | Charset | Id | Default | Compiled | Sortlen |
+--------------------+---------+----+---------+----------+---------+
| latin1_general_cs | latin1 | 49 | | Yes | 1 |
| latin2_czech_cs | latin2 | 2 | | Yes | 4 |
| cp1250_czech_cs | cp1250 | 34 | | Yes | 2 |
| latin7_estonian_cs | latin7 | 20 | | Yes | 1 |
| latin7_general_cs | latin7 | 42 | | Yes | 1 |
| cp1251_general_cs | cp1251 | 52 | | Yes | 1 |
+--------------------+---------+----+---------+----------+---------+
6 rows in set (0.01 sec)
Not much of them on my Debian MySQL default install.
And surprisingly enough, no 'utf8_cs' !
But MySQL has _bin
collation too. Not exactly useful for ordering but quite sufficient for searching:
mysql> show collation where collation like '%_bin';
+--------------+----------+----+---------+----------+---------+
| Collation | Charset | Id | Default | Compiled | Sortlen |
+--------------+----------+----+---------+----------+---------+
| big5_bin | big5 | 84 | | Yes | 1 |
| dec8_bin | dec8 | 69 | | Yes | 1 |
| cp850_bin | cp850 | 80 | | Yes | 1 |
| hp8_bin | hp8 | 72 | | Yes | 1 |
| koi8r_bin | koi8r | 74 | | Yes | 1 |
| latin1_bin | latin1 | 47 | | Yes | 1 |
| latin2_bin | latin2 | 77 | | Yes | 1 |
| swe7_bin | swe7 | 82 | | Yes | 1 |
| ascii_bin | ascii | 65 | | Yes | 1 |
| ujis_bin | ujis | 91 | | Yes | 1 |
| sjis_bin | sjis | 88 | | Yes | 1 |
| hebrew_bin | hebrew | 71 | | Yes | 1 |
| tis620_bin | tis620 | 89 | | Yes | 1 |
| euckr_bin | euckr | 85 | | Yes | 1 |
| koi8u_bin | koi8u | 75 | | Yes | 1 |
| gb2312_bin | gb2312 | 86 | | Yes | 1 |
| greek_bin | greek | 70 | | Yes | 1 |
| cp1250_bin | cp1250 | 66 | | Yes | 1 |
| gbk_bin | gbk | 87 | | Yes | 1 |
| latin5_bin | latin5 | 78 | | Yes | 1 |
| armscii8_bin | armscii8 | 64 | | Yes | 1 |
| utf8_bin | utf8 | 83 | | Yes | 1 |
| ucs2_bin | ucs2 | 90 | | Yes | 1 |
| cp866_bin | cp866 | 68 | | Yes | 1 |
| keybcs2_bin | keybcs2 | 73 | | Yes | 1 |
| macce_bin | macce | 43 | | Yes | 1 |
| macroman_bin | macroman | 53 | | Yes | 1 |
| cp852_bin | cp852 | 81 | | Yes | 1 |
| latin7_bin | latin7 | 79 | | Yes | 1 |
| cp1251_bin | cp1251 | 50 | | Yes | 1 |
| cp1256_bin | cp1256 | 67 | | Yes | 1 |
| cp1257_bin | cp1257 | 58 | | Yes | 1 |
| geostd8_bin | geostd8 | 93 | | Yes | 1 |
| cp932_bin | cp932 | 96 | | Yes | 1 |
| eucjpms_bin | eucjpms | 98 | | Yes | 1 |
+--------------+----------+----+---------+----------+---------+
35 rows in set (0.00 sec)
In the following example, I create a table using utf8 encoding and utf8_bin collation. Please note you could change encoding/collation on a per-column basis instead.
mysql> create TABLE user (username CHAR(20), password CHAR(20)) DEFAULT CHARSET=utf8 DEFAULT COLLATE=utf8_bin;
Query OK, 0 rows affected (0.26 sec)
mysql> insert into user value("ÉleCTeur", "ÉleCTeur");
Query OK, 1 row affected (0.00 sec)
mysql> select count(*) from user where username = "ÉleCTeur" and password="ÉleCTeur";
+----------+
| count(*) |
+----------+
| 1 |
+----------+
1 row in set (0.00 sec)
mysql> select count(*) from user where username = "éleCTeur" and password="éleCTeur";
+----------+
| count(*) |
+----------+
| 0 |
+----------+
1 row in set (0.01 sec)
Upvotes: 2