.. currentmodule:: pandas
.. ipython:: python
:suppress:
import numpy as np
np.random.seed(123456)
from pandas import *
randn = np.random.randn
np.set_printoptions(precision=4, suppress=True)
import matplotlib.pyplot as plt
plt.close('all')
By "group by" we are refer to a process involving one or more of the following steps
- Splitting the data into groups based on some criteria
- Applying a function to each group independently
- Combining the results into a data structure
Of these, the split step is the most straightforward. In fact, in many situations you may wish to split the data set into groups and do something with those groups yourself. In the apply step, we might wish to one of the following:
Aggregation: computing a summary statistic (or statistics) about each group. Some examples:
- Compute group sums or means
- Compute group sizes / counts
Transformation: perform some group-specific computations and return a like-indexed. Some examples:
- Standardizing data (zscore) within group
- Filling NAs within groups with a value derived from each group
Some combination of the above: GroupBy will examine the results of the apply step and try to return a sensibly combined result if it doesn't fit into either of the above two categories
Since the set of object instance method on pandas data structures are generally
rich and expressive, we often simply want to invoke, say, a DataFrame function
on each group. The name GroupBy should be quite familiar to those who have used
a SQL-based tool (or itertools), in which you can write code like:
SELECT Column1, Column2, mean(Column3), sum(Column4)
FROM SomeTable
GROUP BY Column1, Column2We aim to make operations like this natural and easy to express using pandas. We'll address each area of GroupBy functionality then provide some non-trivial examples / use cases.
pandas objects can be split on any of their axes. The abstract definition of grouping is to provide a mapping of labels to group names. To create a GroupBy object (more on what the GroupBy object is later), you do the following:
# default is axis=0
>>> grouped = obj.groupby(key)
>>> grouped = obj.groupby(key, axis=1)
>>> grouped = obj.groupby([key1, key2])
The mapping can be specified many different ways:
- A Python function, to be called on each of the axis labels
- A list or NumPy array of the same length as the selected axis
- A dict or Series, providing a
label -> group namemapping- For DataFrame objects, a string indicating a column to be used to group. Of course
df.groupby('A')is just syntactic sugar fordf.groupby(df['A']), but it makes life simpler- A list of any of the above things
Collectively we refer to the grouping objects as the keys. For example, consider the following DataFrame:
.. ipython:: python
df = DataFrame({'A' : ['foo', 'bar', 'foo', 'bar',
'foo', 'bar', 'foo', 'foo'],
'B' : ['one', 'one', 'two', 'three',
'two', 'two', 'one', 'three'],
'C' : randn(8), 'D' : randn(8)})
df
We could naturally group by either the A or B columns or both:
.. ipython:: python
grouped = df.groupby('A')
grouped = df.groupby(['A', 'B'])
These will split the DataFrame on its index (rows). We could also split by the columns:
.. ipython::
In [4]: def get_letter_type(letter):
...: if letter.lower() in 'aeiou':
...: return 'vowel'
...: else:
...: return 'consonant'
...:
In [5]: grouped = df.groupby(get_letter_type, axis=1)
Starting with 0.8, pandas Index objects now supports duplicate values. If a non-unique index is used as the group key in a groupby operation, all values for the same index value will be considered to be in one group and thus the output of aggregation functions will only contain unique index values:
.. ipython:: python lst = [1, 2, 3, 1, 2, 3] s = Series([1, 2, 3, 10, 20, 30], lst) grouped = s.groupby(level=0) grouped.first() grouped.last() grouped.sum()
Note that no splitting occurs until it's needed. Creating the GroupBy object only verifies that you've passed a valid mapping.
Note
Many kinds of complicated data manipulations can be expressed in terms of GroupBy operations (though can't be guaranteed to be the most efficient). You can get quite creative with the label mapping functions.
The groups attribute is a dict whose keys are the computed unique groups
and corresponding values being the axis labels belonging to each group. In the
above example we have:
.. ipython:: python
df.groupby('A').groups
df.groupby(get_letter_type, axis=1).groups
Calling the standard Python len function on the GroupBy object just returns
the length of the groups dict, so it is largely just a convenience:
.. ipython:: python grouped = df.groupby(['A', 'B']) grouped.groups len(grouped)
By default the group keys are sorted during the groupby operation. You may
however pass sort``=``False for potential speedups:
.. ipython:: python
df2 = DataFrame({'X' : ['B', 'B', 'A', 'A'], 'Y' : [1, 2, 3, 4]})
df2.groupby(['X'], sort=True).sum()
df2.groupby(['X'], sort=False).sum()
With :ref:`hierarchically-indexed data <indexing.hierarchical>`, it's quite natural to group by one of the levels of the hierarchy.
.. ipython:: python
:suppress:
arrays = [['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'],
['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]
tuples = zip(*arrays)
tuples
index = MultiIndex.from_tuples(tuples, names=['first', 'second'])
s = Series(randn(8), index=index)
.. ipython:: python s grouped = s.groupby(level=0) grouped.sum()
If the MultiIndex has names specified, these can be passed instead of the level number:
.. ipython:: python s.groupby(level='second').sum()
The aggregation functions such as sum will take the level parameter
directly. Additionally, the resulting index will be named according to the
chosen level:
.. ipython:: python s.sum(level='second')
Also as of v0.6, grouping with multiple levels is supported.
.. ipython:: python
:suppress:
arrays = [['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'],
['doo', 'doo', 'bee', 'bee', 'bop', 'bop', 'bop', 'bop'],
['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]
tuples = zip(*arrays)
index = MultiIndex.from_tuples(tuples, names=['first', 'second', 'third'])
s = Series(randn(8), index=index)
.. ipython:: python s s.groupby(level=['first','second']).sum()
More on the sum function and aggregation later.
Once you have created the GroupBy object from a DataFrame, for example, you
might want to do something different for each of the columns. Thus, using
[] similar to getting a column from a DataFrame, you can do:
.. ipython:: python grouped = df.groupby(['A']) grouped_C = grouped['C'] grouped_D = grouped['D']
This is mainly syntactic sugar for the alternative and much more verbose:
.. ipython:: python df['C'].groupby(df['A'])
Additionally this method avoids recomputing the internal grouping information derived from the passed key.
With the GroupBy object in hand, iterating through the grouped data is very
natural and functions similarly to itertools.groupby:
.. ipython::
In [4]: grouped = df.groupby('A')
In [5]: for name, group in grouped:
...: print name
...: print group
...:
In the case of grouping by multiple keys, the group name will be a tuple:
.. ipython::
In [5]: for name, group in df.groupby(['A', 'B']):
...: print name
...: print group
...:
It's standard Python-fu but remember you can unpack the tuple in the for loop
statement if you wish: for (k1, k2), group in grouped:.
Once the GroupBy object has been created, several methods are available to
perform a computation on the grouped data. An obvious one is aggregation via
the aggregate or equivalently agg method:
.. ipython:: python
grouped = df.groupby('A')
grouped.aggregate(np.sum)
grouped = df.groupby(['A', 'B'])
grouped.aggregate(np.sum)
As you can see, the result of the aggregation will have the group names as the
new index along the grouped axis. In the case of multiple keys, the result is a
:ref:`MultiIndex <indexing.hierarchical>` by default, though this can be
changed by using the as_index option:
.. ipython:: python
grouped = df.groupby(['A', 'B'], as_index=False)
grouped.aggregate(np.sum)
df.groupby('A', as_index=False).sum()
Note that you could use the reset_index DataFrame function to achieve the
same result as the column names are stored in the resulting MultiIndex:
.. ipython:: python df.groupby(['A', 'B']).sum().reset_index()
Another simple aggregation example is to compute the size of each group.
This is included in GroupBy as the size method. It returns a Series whose
index are the group names and whose values are the sizes of each group.
.. ipython:: python grouped.size()
With grouped Series you can also pass a list or dict of functions to do aggregation with, outputting a DataFrame:
.. ipython:: python
grouped = df.groupby('A')
grouped['C'].agg([np.sum, np.mean, np.std])
If a dict is passed, the keys will be used to name the columns. Otherwise the function's name (stored in the function object) will be used.
.. ipython:: python
grouped['D'].agg({'result1' : np.sum,
'result2' : np.mean})
On a grouped DataFrame, you can pass a list of functions to apply to each column, which produces an aggregated result with a hierarchical index:
.. ipython:: python grouped.agg([np.sum, np.mean, np.std])
Passing a dict of functions has different behavior by default, see the next section.
By passing a dict to aggregate you can apply a different aggregation to the
columns of a DataFrame:
.. ipython:: python
grouped.agg({'C' : np.sum,
'D' : lambda x: np.std(x, ddof=1)})
The function names can also be strings. In order for a string to be valid it must be either implemented on GroupBy or available via :ref:`dispatching <groupby.dispatch>`:
.. ipython:: python
grouped.agg({'C' : 'sum', 'D' : 'std'})
Some common aggregations, currently only sum, mean, and std, have
optimized Cython implementations:
.. ipython:: python
df.groupby('A').sum()
df.groupby(['A', 'B']).mean()
Of course sum and mean are implemented on pandas objects, so the above
code would work even without the special versions via dispatching (see below).
The transform method returns an object that is indexed the same (same size)
as the one being grouped. Thus, the passed transform function should return a
result that is the same size as the group chunk. For example, suppose we wished
to standardize the data within each group:
.. ipython:: python
index = date_range('10/1/1999', periods=1100)
ts = Series(np.random.normal(0.5, 2, 1100), index)
ts = rolling_mean(ts, 100, 100).dropna()
ts.head()
ts.tail()
key = lambda x: x.year
zscore = lambda x: (x - x.mean()) / x.std()
transformed = ts.groupby(key).transform(zscore)
We would expect the result to now have mean 0 and standard deviation 1 within each group, which we can easily check:
.. ipython:: python # Original Data grouped = ts.groupby(key) grouped.mean() grouped.std() # Transformed Data grouped_trans = transformed.groupby(key) grouped_trans.mean() grouped_trans.std()
We can also visually compare the original and transformed data sets.
.. ipython:: python
compare = DataFrame({'Original': ts, 'Transformed': transformed})
@savefig groupby_transform_plot.png width=4in
compare.plot()
Another common data transform is to replace missing data with the group mean.
.. ipython:: python :suppress: cols = ['A', 'B', 'C'] values = randn(1000, 3) values[np.random.randint(0, 1000, 100), 0] = np.nan values[np.random.randint(0, 1000, 50), 1] = np.nan values[np.random.randint(0, 1000, 200), 2] = np.nan data_df = DataFrame(values, columns=cols)
.. ipython:: python data_df countries = np.array(['US', 'UK', 'GR', 'JP']) key = countries[np.random.randint(0, 4, 1000)] grouped = data_df.groupby(key) # Non-NA count in each group grouped.count() f = lambda x: x.fillna(x.mean()) transformed = grouped.transform(f)
We can verify that the group means have not changed in the transformed data and that the transformed data contains no NAs.
.. ipython:: python grouped_trans = transformed.groupby(key) grouped.mean() # original group means grouped_trans.mean() # transformation did not change group means grouped.count() # original has some missing data points grouped_trans.count() # counts after transformation grouped_trans.size() # Verify non-NA count equals group size
When doing an aggregation or transformation, you might just want to call an instance method on each data group. This is pretty easy to do by passing lambda functions:
.. ipython:: python
grouped = df.groupby('A')
grouped.agg(lambda x: x.std())
But, it's rather verbose and can be untidy if you need to pass additional arguments. Using a bit of metaprogramming cleverness, GroupBy now has the ability to "dispatch" method calls to the groups:
.. ipython:: python grouped.std()
What is actually happening here is that a function wrapper is being
generated. When invoked, it takes any passed arguments and invokes the function
with any arguments on each group (in the above example, the std
function). The results are then combined together much in the style of agg
and transform (it actually uses apply to infer the gluing, documented
next). This enables some operations to be carried out rather succinctly:
.. ipython:: python
tsdf = DataFrame(randn(1000, 3),
index=date_range('1/1/2000', periods=1000),
columns=['A', 'B', 'C'])
tsdf.ix[::2] = np.nan
grouped = tsdf.groupby(lambda x: x.year)
grouped.fillna(method='pad')
In this example, we chopped the collection of time series into yearly chunks then independently called :ref:`fillna <missing_data.fillna>` on the groups.
Some operations on the grouped data might not fit into either the aggregate or
transform categories. Or, you may simply want GroupBy to infer how to combine
the results. For these, use the apply function, which can be substituted
for both aggregate and transform in many standard use cases. However,
apply can handle some exceptional use cases, for example:
.. ipython:: python
df
grouped = df.groupby('A')
# could also just call .describe()
grouped['C'].apply(lambda x: x.describe())
The dimension of the returned result can also change:
.. ipython::
In [8]: grouped = df.groupby('A')['C']
In [10]: def f(group):
....: return DataFrame({'original' : group,
....: 'demeaned' : group - group.mean()})
....:
In [11]: grouped.apply(f)
Again consider the example DataFrame we've been looking at:
.. ipython:: python df
Supposed we wished to compute the standard deviation grouped by the A
column. There is a slight problem, namely that we don't care about the data in
column B. We refer to this as a "nuisance" column. If the passed
aggregation function can't be applied to some columns, the troublesome columns
will be (silently) dropped. Thus, this does not pose any problems:
.. ipython:: python
df.groupby('A').std()
If there are any NaN values in the grouping key, these will be automatically excluded. So there will never be an "NA group". This was not the case in older versions of pandas, but users were generally discarding the NA group anyway (and supporting it was an implementation headache).
Categorical variables represented as instance of pandas's Factor class can
be used as group keys. If so, the order of the levels will be preserved:
.. ipython:: python data = Series(np.random.randn(100)) factor = qcut(data, [0, .25, .5, .75, 1.]) data.groupby(factor).mean()