Pandas read_sql() – AttributeError: 'Engine' object has no attribute 'cursor'

Question

I am trying to read data from MySQL query using pandas read_sql() method with python3+sqlalchemy+pymysql

I tried to follow the following tutorials –

https://pythondata.com/quick-tip-sqlalchemy-for-mysql-and-pandas/

https://www.youtube.com/watch?v=M-4EpNdlSuY

https://www.programcreek.com/python/example/101381/pandas.read_sql

Everything just looks good with the code

import pandas
import sqlalchemy
engine = sqlalchemy.create_engine('mysql+pymysql://root:[email protected]:3306/mydatabase')
df = pandas.read_sql("SELECT * FROM persons", con = engine)

Receiving following error –

AttributeError: ‘Engine’ object has no attribute ‘cursor’

When I tried to pass the ‘connection’ variable instead of the ‘engine’, like below –

import pandas
import sqlalchemy
engine = sqlalchemy.create_engine('mysql+pymysql://root:[email protected]:3306/mydatabase')
connection = engine.connect()
df = pandas.read_sql("SELECT * FROM persons", con = connection)

it says –

AttributeError: ‘Connection’ object has no attribute ‘cursor’

What am I doing wrong?

Asked By: Adi

||

Source

Answer 1

I think you are looking for pyodbc and sqlalchemy is another way of query db , check with https://towardsdatascience.com/sqlalchemy-python-tutorial-79a577141a91

import sqlalchemy as sq
engine = sq.create_engine('mysql+pymysql://root:[email protected]:3306/mydatabase')
connection = engine.connect()
metadata = sq.MetaData()
persons = sq.Table('persons', metadata, autoload=True, autoload_with=engine)

Ret = connection.execute(sq.select([persons]))
youdf=pd.DataFrame(Ret.fetchall())

Answered By: BENY

Answer 2

Reference – https://github.com/pandas-dev/pandas/issues/23030#issuecomment-428140488

The problem went away when I

Saved my workbook
Restarted the PC
Reloaded

It seems like something was cached. I could use read_sql() and other pandas sql functions by passing ‘engine’ reference –

import pandas
import sqlalchemy
engine = sqlalchemy.create_engine('mysql+pymysql://root:[email protected]:3306/mydatabase')
df = pandas.read_sql("SELECT * FROM persons", con = engine)

Answered By: Adi

Answer 3

This may be due to a version mismatched. I had a similar error with an old notebook which used to work fine :

AttributeError: ‘Engine’ object has no attribute ‘execution_options’

After a few tests I figured the issue appeared with pandas 1.1.0. In the release notes the minimum version for SQLAlchemy was 1.1.4. Upgrading SQLAlchemy to this version fixed the issue.

Answered By: Skippy le Grand Gourou

Answer 4

I am using MySQL on Win10 with Python 3.10 and Alchemy 1.4

I tried a variety of options nada. THey all blew up, so here is a work around that seems to make everyone happy.

` metadata_obj = MetaData()
tbl_test= Table("test", metadata_obj, autoload_with=engine)

with engine.connect() as sql_con:
    qry = select(tbl_test)
    resultset = sql_con.execute(qry)
    results_as_dict = resultset.mappings().all()
    df = pd.DataFrame(results_as_dict)
    pprint(df)

`

Answered By: QuentinJS

Pandas read_sql() – AttributeError: 'Engine' object has no attribute 'cursor'

Question:

Answers: