IPython notes

This is a post to jog my memory when needed.

IPython saves all it’s hisotry in a SQLite database. This is super useful for recovering commnads from an interactive programming session.

Imagine you forgot to save some dataset manipulation (loading tabdeliminated csv into a DataFrame), you can easily recover that analysis:

Find the command which used the Pandas read_table

In [1]: %history -g read_table
173/5: df = pd.io.parsers.read_table(fname, header=6)
173/9: pd.io.parsers.read_table?
173/10: df = pd.io.parsers.read_table(fname, header=6, index_col=0)
173/12: df = pd.io.parsers.read_table(fname, header=6, index_col=0, parse_dates=True)

Export a chunk of that history to a file for editing:

In [2]: %history 173/12-2000 -f hist.py

Then after editing the file hist.py run this inside your currency IPython session with run -i hist.py (where the -i indicated running the script inside the current sessions namespace).

Sweeet!