Numerical Computing with Python

How it works...

Step 1 formally creates our groupby object. It is useful to display all the public attributes and methods to reveal all the possible functionality as was done in step 2. Each group is uniquely identified by a tuple containing a unique combination of the values in the grouping columns. Pandas allows you to select a specific group as a DataFrame with the get_group method shown in step 5.

It is rare that you will need to iterate through your groups and in general, you should avoid doing so if necessary, as it can be quite slow. Occasionally, you will have no other choice. When iterating through a groupby object, you are given a tuple containing the group name and the DataFrame without the grouping columns. This tuple is unpacked ...