13

How do I encode something in ut8mb4 in Python?

I have two sets of data: data I am migrating to my new MySQL database over from Parse, and data going forward (that talks only to my new database). My database is utf8mb4 in order to store emoji and accented letters.

The first set of data only shows up correctly (when emoji and accents are involved) when I have in my python script:

MySQLdb.escape_string(unicode(xstr(data.get('message'))).encode('utf-8'))

and when reading from the MySQL database in PHP:

$row["message"] = utf8_encode($row["message"]);

The second set of data only shows up correctly (when emoji and accents are involved) when I DON'T include the utf8_encode($row["message"]) portion. I am trying to reconcile these so that both sets of data are returned correctly to my iOS app. Please help!

Improve this question
4
  • 1
    I presume you're talking about MySQL? If so then Python's utf8 should be identical to MySQL's utf8mb4. Can you be more specific about what you're doing and the results you see? Commented Oct 23, 2014 at 16:26
  • It seems utf8mb4 is something MySQL invented because their ut8 implementation only works with BMP characters. This is not the case for normal utf8, which can encode any unicode character. That's the encoding that Python is using. Perhaps MySQL provides a set of functions for working with their custom encoding? Commented Oct 23, 2014 at 16:29
  • @MarkRansom have added more to the question to be more specific Commented Oct 23, 2014 at 17:02
  • It is not clear as to a) what your data looks like in both cases and b) what the data looks like when it doesn't work. Please define clearly what problems you are seeing. Commented Oct 23, 2014 at 18:10

4 Answers 4

Reset to default
28

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.

Improve this answer
Sign up to request clarification or add additional context in comments.

7 Comments

Very useful checklist. Rescued me from a steaming đź’©.
This fixes the issue I was experiencing with mysql + python + emojis đź‘Ź
The coding line applies only to how Python decodes string literals in source code. Not to how encoding and decoding of data is handled. Unless this is Python 2 and you used non-ASCII characters in string literals, there is no need to set this. In Python 3, UTF-8 is the default source encoding.
Beware of this answer if you are on Python3! I've a MySQL database using utf8mb collation. I use MySQLdb with Python3 and I do not need to do anything special to properly read unicode characters. I tried your suggestion and it turned out to perform some sort of double encoding on data, returning some weird unicode characters instead of the right ones.
@baxeico I am not sure if I ever tested my answer in Python 3 back in the day. I know that above was a big issue in Python 2. I'll do some testing in Python 3 and update my answer.
|
27

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.

Improve this answer

Comments

2

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")
Improve this answer

Comments

2

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.

Improve this answer

Comments

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.