Dataset is a Python library, which is basically a wrapper around SQLAlchemy. It claims to be so easy to use that even lazy people like it.

Install dataset as follows:

$ sudo pip install dataset
$ pip freeze|grep dataset
dataset==0.5.4

Create a SQLite in-memory database and connect to it:

import dataset
db = dataset.connect('sqlite:///:memory:')

Create a table called books:

table = db["books"]

Actually, the table in the database isn't created yet, since we haven't specified any columns. We only created a related object. The table schema is created automatically from calls to the insert() method. Give the insert() method dictionaries with book titles:

table.insert(dict(title="NumPy Beginner's Guide", author='Ivan Idris'))
table.insert(dict(title="NumPy Cookbook", author='Ivan Idris'))
table.insert(dict(title="Learning NumPy", author='Ivan Idris'))

These are all excellent books, of course! The read_sql() pandas function can query this table too:

print read_sql(...