.. currentmodule:: pandas
.. ipython:: python
:suppress:
import numpy as np
np.random.seed(123456)
np.set_printoptions(precision=4, suppress=True)
import pandas as pd
pd.options.display.max_rows = 15
import matplotlib
# matplotlib.style.use('default')
import matplotlib.pyplot as plt
plt.close('all')
from collections import OrderedDict
Split-apply-combine is a common paradigm in data analysis. It involves splitting
the data set into smaller groups, applying some operation to each group
independently and combining the results into a data structure. This strategy is
supported for example in Excels pivot tables, SQLs group by operator and Rs
plyr package. This section will look at the the Pandas groupby and
related functions and show you how to do split-apply-combine in Pandas. See the
:ref:`cookbook<cookbook.grouping>` for some advanced strategies
The split step is the most straightforward. See the section on :ref:`splitting<groupby.split>` below.
- In the apply step you may wish to apply one of the following operations:
- Aggregate: Get a single value for each group. This could be a summary statistic like sum or mean of some column or counting the number of members in the group. See the section on :ref:`aggregating<groupby.aggregate>` below.
- Filter: When you want a subset of your original data. Discard data according to some function applied to each group. This can be useful when for example you wish to discard groups with low member count. See the section on :ref:`filtering<groupby.filter>` below.
- Transform: A new value for each original row. This can be used to normalize/scale data or filling in erroneous or missing values. See the section on :ref:`transforming<groupby.transform>` below.
Pandas has direct support for these three operations and will try and return a sensibly combined result. See here for further help on when to use aggregate/filter/transform in Pandas.
Pandas also supports iteration over the groups created in the split step; Using iteration over the groups (rather than the three shortcut functions) renders more control over the apply and combine parts of the process, but also requires more work from the programmer. See the section in :ref:`iterating<groupby.iterating>` below.
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- For DataFrame objects, a string indicating an index level to be used to group.
- A list of any of the above things
Collectively we refer to the grouping objects as the keys. For example, consider the following DataFrame:
Note
.. versionadded:: 0.20
A string passed to groupby may refer to either a column or an index level.
If a string matches both a column name and an index level name then a warning is
issued and the column takes precedence. This will result in an ambiguity error
in a future version.
.. ipython:: python
df = pd.DataFrame({'A' : ['foo', 'bar', 'foo', 'bar',
'foo', 'bar', 'foo', 'foo'],
'B' : ['one', 'one', 'two', 'three',
'two', 'two', 'one', 'three'],
'C' : np.random.randn(8),
'D' : np.random.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)
pandas Index objects support 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 = pd.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.
By default the group keys are sorted during the groupby operation. You may however pass sort=False for potential speedups:
.. ipython:: python
df2 = pd.DataFrame({'X' : ['B', 'B', 'A', 'A'], 'Y' : [1, 2, 3, 4]})
df2.groupby(['X']).sum()
df2.groupby(['X'], sort=False).sum()
Note that groupby will preserve the order in which observations are sorted within each group.
For example, the groups created by groupby() below are in the order they appeared in the original DataFrame:
.. ipython:: python
df3 = pd.DataFrame({'X' : ['A', 'B', 'A', 'B'], 'Y' : [1, 4, 3, 2]})
df3.groupby(['X']).get_group('A')
df3.groupby(['X']).get_group('B')
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)
GroupBy will tab complete column names (and other attributes)
.. ipython:: python
:suppress:
n = 10
weight = np.random.normal(166, 20, size=n)
height = np.random.normal(60, 10, size=n)
time = pd.date_range('1/1/2000', periods=n)
gender = np.random.choice(['male', 'female'], size=n)
df = pd.DataFrame({'height': height, 'weight': weight,
'gender': gender}, index=time)
.. ipython:: python
df
gb = df.groupby('gender')
.. ipython:: @verbatim In [1]: gb.<TAB> gb.agg gb.boxplot gb.cummin gb.describe gb.filter gb.get_group gb.height gb.last gb.median gb.ngroups gb.plot gb.rank gb.std gb.transform gb.aggregate gb.count gb.cumprod gb.dtype gb.first gb.groups gb.hist gb.max gb.min gb.nth gb.prod gb.resample gb.sum gb.var gb.apply gb.cummax gb.cumsum gb.fillna gb.gender gb.head gb.indices gb.mean gb.name gb.ohlc gb.quantile gb.size gb.tail gb.weight
With :ref:`hierarchically-indexed data <advanced.hierarchical>`, it's quite natural to group by one of the levels of the hierarchy.
Let's create a Series with a two-level MultiIndex.
.. ipython:: python
arrays = [['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'],
['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]
index = pd.MultiIndex.from_arrays(arrays, names=['first', 'second'])
s = pd.Series(np.random.randn(8), index=index)
s
We can then group by one of the levels in s.
.. ipython:: python 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')
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 = list(zip(*arrays))
index = pd.MultiIndex.from_tuples(tuples, names=['first', 'second', 'third'])
s = pd.Series(np.random.randn(8), index=index)
.. ipython:: python s s.groupby(level=['first', 'second']).sum()
.. versionadded:: 0.20
Index level names may be supplied as keys.
.. ipython:: python s.groupby(['first', 'second']).sum()
More on the sum function and aggregation later.
A DataFrame may be grouped by a combination of columns and index levels by
specifying the column names as strings and the index levels as pd.Grouper
objects.
.. ipython:: python
arrays = [['bar', 'bar', 'baz', 'baz', 'foo', 'foo', 'qux', 'qux'],
['one', 'two', 'one', 'two', 'one', 'two', 'one', 'two']]
index = pd.MultiIndex.from_arrays(arrays, names=['first', 'second'])
df = pd.DataFrame({'A': [1, 1, 1, 1, 2, 2, 3, 3],
'B': np.arange(8)},
index=index)
df
The following example groups df by the second index level and
the A column.
.. ipython:: python df.groupby([pd.Grouper(level=1), 'A']).sum()
Index levels may also be specified by name.
.. ipython:: python df.groupby([pd.Grouper(level='second'), 'A']).sum()
.. versionadded:: 0.20
Index level names may be specified as keys directly to groupby.
.. ipython:: python df.groupby(['second', 'A']).sum()
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
:suppress:
df = pd.DataFrame({'A' : ['foo', 'bar', 'foo', 'bar',
'foo', 'bar', 'foo', 'foo'],
'B' : ['one', 'one', 'two', 'three',
'two', 'two', 'one', 'three'],
'C' : np.random.randn(8),
'D' : np.random.randn(8)})
.. 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:.
A single group can be selected using GroupBy.get_group():
.. ipython:: python
grouped.get_group('bar')
Or for an object grouped on multiple columns:
.. ipython:: python
df.groupby(['A', 'B']).get_group(('bar', 'one'))
This section describes how to aggregate data. We will be giving examples using
the tips.csv dataset. Each row represents a meal at some restaurant; The
columns store the value of the total bill, the size if the tip and some metadata
about the customer.
.. ipython:: python
tips = pd.read_csv('./data/tips.csv')
tips
What if we wanted to know the average total bill on each day? We split the data so that each group consists of all the meals eaten on the same day. We want a single value for each group, so we should use the aggregate function:
.. ipython:: python
tips.groupby('day').aggregate('mean')
The result has the group names, in this case the days, as the index along the
grouped axis. Along the other axis we have the columns for which Pandas could
calculate a mean, i.e. the ones with a numeric data type. We could have selected
the total_bill column either before or after aggregating to limit the result
to this column, but not before splitting since we need the day column for
the splitting.
How about the number of guests for each day and for each time of day? In this
case it is not enough to split the data on the day it was eaten, we also need
split by the time of day. Instead of calculating the mean, like in the previous
example we use the sum function.
.. ipython:: python
tips.groupby(['day', 'time'])['size'].agg('sum')
agg is short for aggregate.
Pandas has support for a number of basic descriptive statistic functions which can be used with aggregate:
| Function | Description |
|---|---|
count |
Number of non-NA observations |
sum |
Sum of values |
mean |
Mean of values |
mad |
Mean absolute deviation |
median |
Arithmetic median of values |
min |
Minimum |
max |
Maximum |
mode |
Mode |
abs |
Absolute Value |
prod |
Product of values |
std |
Bessel-corrected sample standard deviation |
var |
Unbiased variance |
sem |
Standard error of the mean |
skew |
Sample skewness (3rd moment) |
kurt |
Sample kurtosis (4th moment) |
quantile |
Sample quantile (value at %) |
What if we need to know the difference between the smallest and largest total
bill for each day? Again we split the data so each group has the meals eaten on
the same day. But which function do we use to find the difference? The agg
function also accepts a function as argument. The function is called once per
group with the current group as argument and should return a single value.
.. ipython:: python tips.groupby(['size']).agg(lambda group: max(group) - min(group))['total_bill']
Once the GroupBy object has been created, several methods are available to perform a computation on the grouped data. These operations are similar to the :ref:`aggregating API <basics.aggregate>`, :ref:`window functions API <stats.aggregate>`, and :ref:`resample API <timeseries.aggregate>`.
Note
Aggregation functions will not return the groups that you are aggregating over
if they are named columns, when as_index=True, the default. The grouped columns will
be the indices of the returned object.
Passing as_index=False will return the groups that you are aggregating over, if they are
named columns.
Aggregating functions are ones that reduce the dimension of the returned objects,
for example: mean, sum, size, count, std, var, sem, describe, first, last, nth, min, max. This is
what happens when you do for example DataFrame.sum() and get back a Series.
nth can act as a reducer or a filter, see :ref:`here <groupby.nth>`
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])
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])
The resulting aggregations are named for the functions themselves. If you
need to rename, then you can add in a chained operation for a Series like this:
.. ipython:: python
(grouped['C'].agg([np.sum, np.mean, np.std])
.rename(columns={'sum': 'foo',
'mean': 'bar',
'std': 'baz'})
)
For a grouped DataFrame, you can rename in a similar manner:
.. ipython:: python
(grouped.agg([np.sum, np.mean, np.std])
.rename(columns={'sum': 'foo',
'mean': 'bar',
'std': 'baz'})
)
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'})
Note
If you pass a dict to aggregate, the ordering of the output columns is
non-deterministic. If you want to be sure the output columns will be in a specific
order, you can use an OrderedDict. Compare the output of the following two commands:
.. ipython:: python
grouped.agg({'D': 'std', 'C': 'mean'})
grouped.agg(OrderedDict([('D', 'std'), ('C', 'mean')]))
Some common aggregations, currently only sum, mean, std, and sem, 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. The transform function must:
- Return a result that is either the same size as the group chunk or
broadcastable to the size of the group chunk (e.g., a scalar,
grouped.transform(lambda x: x.iloc[-1])). - Operate column-by-column on the group chunk. The transform is applied to the first group chunk using chunk.apply.
- Not perform in-place operations on the group chunk. Group chunks should
be treated as immutable, and changes to a group chunk may produce unexpected
results. For example, when using
fillna,inplacemust beFalse(grouped.transform(lambda x: x.fillna(inplace=False))). - (Optionally) operates on the entire group chunk. If this is supported, a fast path is used starting from the second chunk.
For example, suppose we wished to standardize the data within each group:
.. ipython:: python
index = pd.date_range('10/1/1999', periods=1100)
ts = pd.Series(np.random.normal(0.5, 2, 1100), index)
ts = ts.rolling(window=100,min_periods=100).mean().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 = pd.DataFrame({'Original': ts, 'Transformed': transformed})
@savefig groupby_transform_plot.png
compare.plot()
Transformation functions that have lower dimension outputs are broadcast to match the shape of the input array.
.. ipython:: python data_range = lambda x: x.max() - x.min() ts.groupby(key).transform(data_range)
Alternatively the built-in methods can be could be used to produce the same outputs
.. ipython:: python
ts.groupby(key).transform('max') - ts.groupby(key).transform('min')
Another common data transform is to replace missing data with the group mean.
.. ipython:: python :suppress: cols = ['A', 'B', 'C'] values = np.random.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 = pd.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
Note
Some functions when applied to a groupby object will automatically transform
the input, returning an object of the same shape as the original. Passing
as_index=False will not affect these transformation methods.
For example: fillna, ffill, bfill, shift.
.. ipython:: python grouped.ffill()
.. versionadded:: 0.18.1
Working with the resample, expanding or rolling operations on the groupby
level used to require the application of helper functions. However,
now it is possible to use resample(), expanding() and
rolling() as methods on groupbys.
The example below will apply the rolling() method on the samples of
the column B based on the groups of column A.
.. ipython:: python
df_re = pd.DataFrame({'A': [1] * 10 + [5] * 10,
'B': np.arange(20)})
df_re
df_re.groupby('A').rolling(4).B.mean()
The expanding() method will accumulate a given operation
(sum() in the example) for all the members of each particular
group.
.. ipython:: python
df_re.groupby('A').expanding().sum()
Suppose you want to use the resample() method to get a daily
frequency in each group of your dataframe and wish to complete the
missing values with the ffill() method.
.. ipython:: python
df_re = pd.DataFrame({'date': pd.date_range(start='2016-01-01',
periods=4,
freq='W'),
'group': [1, 1, 2, 2],
'val': [5, 6, 7, 8]}).set_index('date')
df_re
df_re.groupby('group').resample('1D').ffill()
The filter method returns a subset of the original object. Suppose we
want to take only elements that belong to groups with a group sum greater
than 2.
.. ipython:: python sf = pd.Series([1, 1, 2, 3, 3, 3]) sf.groupby(sf).filter(lambda x: x.sum() > 2)
The argument of filter must be a function that, applied to the group as a
whole, returns True or False.
Another useful operation is filtering out elements that belong to groups with only a couple members.
.. ipython:: python
dff = pd.DataFrame({'A': np.arange(8), 'B': list('aabbbbcc')})
dff.groupby('B').filter(lambda x: len(x) > 2)
Alternatively, instead of dropping the offending groups, we can return a like-indexed objects where the groups that do not pass the filter are filled with NaNs.
.. ipython:: python
dff.groupby('B').filter(lambda x: len(x) > 2, dropna=False)
For DataFrames with multiple columns, filters should explicitly specify a column as the filter criterion.
.. ipython:: python
dff['C'] = np.arange(8)
dff.groupby('B').filter(lambda x: len(x['C']) > 2)
Note
Some functions when applied to a groupby object will act as a filter on the input, returning
a reduced shape of the original (and potentially eliminating groups), but with the index unchanged.
Passing as_index=False will not affect these transformation methods.
For example: head, tail.
.. ipython:: python
dff.groupby('B').head(2)
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 = pd.DataFrame(np.random.randn(1000, 3),
index=pd.date_range('1/1/2000', periods=1000),
columns=['A', 'B', 'C'])
tsdf.iloc[::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.
The nlargest and nsmallest methods work on Series style groupbys:
.. ipython:: python
s = pd.Series([9, 8, 7, 5, 19, 1, 4.2, 3.3])
g = pd.Series(list('abababab'))
gb = s.groupby(g)
gb.nlargest(3)
gb.nsmallest(3)
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 pd.DataFrame({'original' : group,
....: 'demeaned' : group - group.mean()})
....:
In [11]: grouped.apply(f)
apply on a Series can operate on a returned value from the applied function, that is itself a series, and possibly upcast the result to a DataFrame
.. ipython:: python
def f(x):
return pd.Series([ x, x**2 ], index = ['x', 'x^2'])
s = pd.Series(np.random.rand(5))
s
s.apply(f)
Note
apply can act as a reducer, transformer, or filter function, depending on exactly what is passed to it.
So depending on the path taken, and exactly what you are grouping. Thus the grouped columns(s) may be included in
the output as well as set the indices.
Warning
In the current implementation apply calls func twice on the first group to decide whether it can take a fast or slow code path. This can lead to unexpected behavior if func has side-effects, as they will take effect twice for the first group.
.. ipython:: python
d = pd.DataFrame({"a":["x", "y"], "b":[1,2]})
def identity(df):
print(df)
return df
d.groupby("a").apply(identity)
Again consider the example DataFrame we've been looking at:
.. ipython:: python df
Suppose we wish 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 or NaT values in the grouping key, these will be automatically excluded. So there will never be an "NA group" or "NaT 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 Categorical class
can be used as group keys. If so, the order of the levels will be preserved:
.. ipython:: python data = pd.Series(np.random.randn(100)) factor = pd.qcut(data, [0, .25, .5, .75, 1.]) data.groupby(factor).mean()
You may need to specify a bit more data to properly group. You can
use the pd.Grouper to provide this local control.
.. ipython:: python
import datetime
df = pd.DataFrame({
'Branch' : 'A A A A A A A B'.split(),
'Buyer': 'Carl Mark Carl Carl Joe Joe Joe Carl'.split(),
'Quantity': [1,3,5,1,8,1,9,3],
'Date' : [
datetime.datetime(2013,1,1,13,0),
datetime.datetime(2013,1,1,13,5),
datetime.datetime(2013,10,1,20,0),
datetime.datetime(2013,10,2,10,0),
datetime.datetime(2013,10,1,20,0),
datetime.datetime(2013,10,2,10,0),
datetime.datetime(2013,12,2,12,0),
datetime.datetime(2013,12,2,14,0),
]
})
df
Groupby a specific column with the desired frequency. This is like resampling.
.. ipython:: python df.groupby([pd.Grouper(freq='1M',key='Date'),'Buyer']).sum()
You have an ambiguous specification in that you have a named index and a column that could be potential groupers.
.. ipython:: python
df = df.set_index('Date')
df['Date'] = df.index + pd.offsets.MonthEnd(2)
df.groupby([pd.Grouper(freq='6M',key='Date'),'Buyer']).sum()
df.groupby([pd.Grouper(freq='6M',level='Date'),'Buyer']).sum()
Just like for a DataFrame or Series you can call head and tail on a groupby:
.. ipython:: python
df = pd.DataFrame([[1, 2], [1, 4], [5, 6]], columns=['A', 'B'])
df
g = df.groupby('A')
g.head(1)
g.tail(1)
This shows the first or last n rows from each group.
To select from a DataFrame or Series the nth item, use the nth method. This is a reduction method, and will return a single row (or no row) per group if you pass an int for n:
.. ipython:: python
df = pd.DataFrame([[1, np.nan], [1, 4], [5, 6]], columns=['A', 'B'])
g = df.groupby('A')
g.nth(0)
g.nth(-1)
g.nth(1)
If you want to select the nth not-null item, use the dropna kwarg. For a DataFrame this should be either 'any' or 'all' just like you would pass to dropna:
.. ipython:: python # nth(0) is the same as g.first() g.nth(0, dropna='any') g.first() # nth(-1) is the same as g.last() g.nth(-1, dropna='any') # NaNs denote group exhausted when using dropna g.last() g.B.nth(0, dropna='all')
As with other methods, passing as_index=False, will achieve a filtration, which returns the grouped row.
.. ipython:: python
df = pd.DataFrame([[1, np.nan], [1, 4], [5, 6]], columns=['A', 'B'])
g = df.groupby('A',as_index=False)
g.nth(0)
g.nth(-1)
You can also select multiple rows from each group by specifying multiple nth values as a list of ints.
.. ipython:: python business_dates = pd.date_range(start='4/1/2014', end='6/30/2014', freq='B') df = pd.DataFrame(1, index=business_dates, columns=['a', 'b']) # get the first, 4th, and last date index for each month df.groupby((df.index.year, df.index.month)).nth([0, 3, -1])
To see the order in which each row appears within its group, use the
cumcount method:
.. ipython:: python
dfg = pd.DataFrame(list('aaabba'), columns=['A'])
dfg
dfg.groupby('A').cumcount()
dfg.groupby('A').cumcount(ascending=False)
.. versionadded:: 0.20.2
To see the ordering of the groups (as opposed to the order of rows
within a group given by cumcount) you can use the ngroup
method.
Note that the numbers given to the groups match the order in which the groups would be seen when iterating over the groupby object, not the order they are first observed.
.. ipython:: python
dfg = pd.DataFrame(list('aaabba'), columns=['A'])
dfg
dfg.groupby('A').ngroup()
dfg.groupby('A').ngroup(ascending=False)
Groupby also works with some plotting methods. For example, suppose we suspect that some features in a DataFrame may differ by group, in this case, the values in column 1 where the group is "B" are 3 higher on average.
.. ipython:: python np.random.seed(1234) df = pd.DataFrame(np.random.randn(50, 2)) df['g'] = np.random.choice(['A', 'B'], size=50) df.loc[df['g'] == 'B', 1] += 3
We can easily visualize this with a boxplot:
.. ipython:: python
:okwarning:
@savefig groupby_boxplot.png
df.groupby('g').boxplot()
The result of calling boxplot is a dictionary whose keys are the values
of our grouping column g ("A" and "B"). The values of the resulting dictionary
can be controlled by the return_type keyword of boxplot.
See the :ref:`visualization documentation<visualization.box>` for more.
Warning
For historical reasons, df.groupby("g").boxplot() is not equivalent
to df.boxplot(by="g"). See :ref:`here<visualization.box.return>` for
an explanation.
Regroup columns of a DataFrame according to their sum, and sum the aggregated ones.
.. ipython:: python
df = pd.DataFrame({'a':[1,0,0], 'b':[0,1,0], 'c':[1,0,0], 'd':[2,3,4]})
df
df.groupby(df.sum(), axis=1).sum()
By using .ngroup(), we can extract information about the groups in
a way similar to :func:`factorize` (as described further in the
:ref:`reshaping API <reshaping.factorize>`) but which applies
naturally to multiple columns of mixed type and different
sources. This can be useful as an intermediate categorical-like step
in processing, when the relationships between the group rows are more
important than their content, or as input to an algorithm which only
accepts the integer encoding. (For more information about support in
pandas for full categorical data, see the :ref:`Categorical
introduction <categorical>` and the
:ref:`API documentation <api.categorical>`.)
.. ipython:: python
dfg = pd.DataFrame({"A": [1, 1, 2, 3, 2], "B": list("aaaba")})
dfg
dfg.groupby(["A", "B"]).ngroup()
dfg.groupby(["A", [0, 0, 0, 1, 1]]).ngroup()
Resampling produces new hypothetical samples (resamples) from already existing observed data or from a model that generates data. These new samples are similar to the pre-existing samples.
In order to resample to work on indices that are non-datetimelike, the following procedure can be utilized.
In the following examples, df.index // 5 returns a binary array which is used to determine what gets selected for the groupby operation.
Note
The below example shows how we can downsample by consolidation of samples into fewer samples. Here by using df.index // 5, we are aggregating the samples in bins. By applying std() function, we aggregate the information contained in many samples into a small subset of values which is their standard deviation thereby reducing the number of samples.
.. ipython:: python df = pd.DataFrame(np.random.randn(10,2)) df df.index // 5 df.groupby(df.index // 5).std()
Group DataFrame columns, compute a set of metrics and return a named Series. The Series name is used as the name for the column index. This is especially useful in conjunction with reshaping operations such as stacking in which the column index name will be used as the name of the inserted column:
.. ipython:: python
df = pd.DataFrame({
'a': [0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2],
'b': [0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1],
'c': [1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0],
'd': [0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1],
})
def compute_metrics(x):
result = {'b_sum': x['b'].sum(), 'c_mean': x['c'].mean()}
return pd.Series(result, name='metrics')
result = df.groupby('a').apply(compute_metrics)
result
result.stack()