Lesson 6

Lets take a look at the groupby function.

In [1]:
# Import libraries
from pandas import DataFrame
import pandas as pd
In [2]:
print 'Pandas version: ' + pd.__version__
Pandas version: 0.13.0

In [3]:
# Our small data set
d = {'one':[1,1,1,1,1],
     'two':[2,2,2,2,2],
     'letter':['a','a','b','b','c']}

# Create dataframe
df = DataFrame(d)
df
Out[3]:
letter one two
0 a 1 2
1 a 1 2
2 b 1 2
3 b 1 2
4 c 1 2

5 rows × 3 columns

In [4]:
# Create group object
one = df.groupby('letter')

# Apply sum function
one.sum()
Out[4]:
one two
letter
a 2 4
b 2 4
c 1 2

3 rows × 2 columns

In [5]:
letterone = df.groupby(['letter','one']).sum()
letterone
Out[5]:
two
letter one
a 1 4
b 1 4
c 1 2

3 rows × 1 columns

In [6]:
letterone.index
Out[6]:
MultiIndex(levels=[[u'a', u'b', u'c'], [1]],
           labels=[[0, 1, 2], [0, 0, 0]],
           names=[u'letter', u'one'])

You may want to not have the columns you are grouping by become your index, this can be easily achieved as shown below.

In [7]:
letterone = df.groupby(['letter','one'], as_index=False).sum()
letterone
Out[7]:
letter one two
0 a 1 4
1 b 1 4
2 c 1 2

3 rows × 3 columns

In [8]:
letterone.index
Out[8]:
Int64Index([0, 1, 2], dtype='int64')

Author: David Rojas LLC

Back to top