How to encode (utf8mb4) in Python

I have struggled myself with the correct exchange of the full range of UTF-8 characters between Python and MySQL for the sake of Emoji and other characters beyond the U+FFFF codepoint.

To be sure that everything worked fine, I had to do the following:

1. make sure utf8mb4 was used for CHAR, VARCHAR, and TEXT columns in MySQL
2. enforce UTF-8 in Python
3. enforce UTF-8 to be used between Python and MySQL

To enforce UTF-8 in Python, add the following line as first or second line of your Python script:

# -*- coding: utf-8 -*-

To enforce UTF-8 between Python and MySQL, setup the MySQL connection as follows:

# Connect to mysql.
dbc = MySQLdb.connect(host='###', user='###', passwd='###', db='###', use_unicode=True)

# Create a cursor.
cursor = dbc.cursor()

# Enforce UTF-8 for the connection.
cursor.execute('SET NAMES utf8mb4')
cursor.execute("SET CHARACTER SET utf8mb4")
cursor.execute("SET character_set_connection=utf8mb4")

# Do database stuff.

# Commit data.
dbc.commit()

# Close cursor and connection.
cursor.close()
dbc.close()

This way, you don't need to use functions such as encode and utf8_encode.

use_unicode=True didn't work for me.

My solution

• in mysql, change entire database, table and field encoding to utf8mb4
• MySQLdb.connect(host='###' [...], charset='utf8'
• dbCursor.execute('SET NAMES utf8mb4')
• dbCursor.execute("SET CHARACTER SET utf8mb4")

MySQL's utf8mb4 encoding is just standard UTF-8.

They had to add that name however to distinguish it from the broken UTF-8 character set which only supported BMP characters.

In other words, from the Python side you should always encode to UTF-8 when talking to MySQL, but take into account that the database may not be able to handle Unicode codepoints beyond U+FFFF, unless you use utf8mb4 on the MySQL side.

However, generally speaking, you want to avoid manually encoding and decoding, and instead leave it to MySQLdb worry about this. You do this by configuring your connection and your collations to handle Unicode text transparently. For MySQLdb, that means setting charset='utf8mb4':

database = MySQLdb.connect(
    host=hostname,
    user=username,
    passwd=password,
    db=databasename,
    charset="utf8mb4"
)

Then use normal Python 3 str strings; leave the use_unicode option set to it's default True^*.

Note: this handles SET NAMES and SET character_set_connection) for you, there is no need to issue those manually.

---

^* Unless you still use Python 2, then the default is False. Set it to True and use u'...' unicode strings.

You can also enter the type of code that you want in the following way

mysql.connector.connect(host = '<host>', database = '<db>', user = '<user>', password = '<password>', charset = 'utf8')

The fields inside '<>' are your own details. Instead of 'utf8' you can also write 'utf8mb4' depending on the type of coding your mysqldb wants.