I\'m writing a pandas Dataframe to a Postgres database:
from sqlalchemy import create_engine, MetaData
engine = create_engine(r\'postgresql://user:password@local
I have encountered this error recently and it was solved my removing the .pyc files located in the same directory as .py files. These files (.pyc) holds the previous version information and time, date.
I suppose you are using pandas 0.15. PandasSQLAlchemy
was not yet really public, and was renamed in pandas 0.15 to SQLDatabase
. So if you replace that in your code, it should work (so pdsql = pd.io.sql.SQLDatabase(engine, meta=meta)
).
However, starting from pandas 0.15, there is also schema support in the read_sql_table
and to_sql
functions, so it should not be needed to make a MetaData
and SQLDatabase
object manually. Instead, this should do it:
dataframe.to_sql(table_name, engine, schema='data_quality')
See the 0.15 release notes: http://pandas.pydata.org/pandas-docs/stable/whatsnew.html#improvements-in-the-sql-io-module