dfdb

Pandas-like interface for manipulating SQL databases

Install

pip install dfdb

Overview

This package is a lightweight wrapper around SQLalchemy. As such, it can be used with a variety of database backends, although as of now the package has only been tested with SQLite.

The dfdb.Database class provides a dict-like interface of manipulating tables in a SQL database. For example, if we create a Database object like

db = dfdb.Database(filename='data.db', drivername='sqlite')

Then we can perform operations on a table called df in data.db by calling, for example

db['df'].groupby('col1').agg(mean=('col2', 'mean')).fetch()

db['df'] is a dfdb.DataFrame object that behaves almost identically as a pandas.DataFrame. The only difference is that you need to call .fetch(), which establishes a connection to the database and performs the corresponding SQL query.

Adding rows to the table is as simple as

db['df'].append({'col1': 1, 'col2': 2})

More detailed documentation to follow in the future.