Insert pandas dataframe to mysql using sqlalchemy

Question

I simply try to write a pandas dataframe to local mysql database on ubuntu.

from sqlalchemy import create_engine
import tushare as ts

df = ts.get_tick_data('600848', date='2014-12-22')
engine = create_engine('mysql://user:passwd@127.0.0.1/db_name?charset=utf8')
df.to_sql('tick_data',engine, flavor = 'mysql', if_exists= 'append')

and it pop the error

biggreyhairboy@ubuntu:~/git/python/fjb$ python tushareDB.py 
Error on sql SHOW TABLES LIKE 'tick_data'
Traceback (most recent call last):
 File "tushareDB.py", line 13, in <module>
   df.to_sql('tick_data', con = engine,flavor ='mysql', if_exists= 'append')
  File "/usr/lib/python2.7/dist-packages/pandas/core/frame.py", line 1261, in to_sql
    self, name, con, flavor=flavor, if_exists=if_exists, **kwargs)
  File "/usr/lib/python2.7/dist-packages/pandas/io/sql.py", line 207, in write_frame
    exists = table_exists(name, con, flavor)
  File "/usr/lib/python2.7/dist-packages/pandas/io/sql.py", line 275, in table_exists
    return len(tquery(query, con)) > 0
  File "/usr/lib/python2.7/dist-packages/pandas/io/sql.py", line 90, in tquery
    cur = execute(sql, con, cur=cur)
  File "/usr/lib/python2.7/dist-packages/pandas/io/sql.py", line 53, in execute
    con.rollback()
AttributeError: 'Engine' object has no attribute 'rollback'

the dataframe is not empty, database is ready without tables, i have tried other method to create table in python with mysqldb and it works fine.

a related question: Writing to MySQL database with pandas using SQLAlchemy, to_sql but no actual reason was explained

Answer 1

You appear to be using an older version of pandas. I did a quick git bisect to find the version of pandas where line 53 contains con.rollback() , and found pandas at v0.12, which is before SQLAlchemy support was added to the execute function.

If you're stuck on this version of pandas, you'll need to use a raw DBAPI connection:

df.to_sql('tick_data', engine.raw_connection(), flavor='mysql', if_exists='append')

Otherwise, update pandas and use the engine as you intend to. Note that you don't need to use the flavor parameter when using SQLAlchemy:

df.to_sql('tick_data', engine, if_exists='append')