upload pandas dataframe to redshift - relation "sqlite_master" does not exist

Question

I am trying to write a dataframe from pandas to redshift.

here is the code

df = pd.DataFrame({'num_legs': [2, 4, 8, 0],
                   'num_wings': [2, 0, 0, 0],
                   'num_specimen_seen': [10, 2, 1, 8]},
                  index=['falcon', 'dog', 'spider', 'fish'])

from sqlalchemy import create_engine
import sqlalchemy
sql_engine = create_engine('postgresql://username:password@host:port/dbname')
conn = sql_engine.raw_connection()

df.to_sql('tmp_table', conn, index = False, if_exists = 'replace')

However, I get the following error

    ---------------------------------------------------------------------------
    UndefinedTable                            Traceback (most recent call last)
    ~/opt/anaconda3/envs/UserExperience/lib/python3.7/site-packages/pandas/io/sql.py in execute(self, *args, **kwargs)
       1594             else:
    -> 1595                 cur.execute(*args)
       1596             return cur

    UndefinedTable: relation "sqlite_master" does not exist
...
...
...

   1593                 cur.execute(*args, **kwargs)
   1594             else:
-> 1595                 cur.execute(*args)
   1596             return cur
   1597         except Exception as exc:

DatabaseError: Execution failed on sql 'SELECT name FROM sqlite_master WHERE type='table' AND name=?;': relation "sqlite_master" does not exist

I tried to user pandas_redshift however, seems first one has to upload to s3 bucket and then to the redshift. I would like to directly upload. Similarly, Here I see the answer suggest to upload to s3 first and then to the redshift

I can read and do query on the database using the same connection.

Answer 1

Try using sql_engine instead of conn.

I just had the same issue and using engine did the trick, try the following:

import sqlalchemy

engine = sqlalchemy.create_engine('postgres://username:password@url:5439/db_name')
print(bool(engine)) # <- just to keep track of the process

with engine.connect() as conn:
    print(bool(conn)) # <- just to keep track of the process

    df.to_sql(name=table_name, con=engine)
    print("end") # <- just to keep track of the process