Matching / broadcasting behavior¶

DataFrame has the methods add(), sub(),
mul(), div() and related functions
radd(), rsub(), …
for carrying out binary operations. For broadcasting behavior,
Series input is of primary interest. Using these functions, you can use to
either match on the index or columns via the axis keyword:

In [18]: df = pd.DataFrame({
   ....:     'one': pd.Series(np.random.randn(3), index=['a', 'b', 'c']),
   ....:     'two': pd.Series(np.random.randn(4), index=['a', 'b', 'c', 'd']),
   ....:     'three': pd.Series(np.random.randn(3), index=['b', 'c', 'd'])})
   ....: 

In [19]: df
Out[19]: 
        one       two     three
a  1.394981  1.772517       NaN
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [20]: row = df.iloc[1]

In [21]: column = df['two']

In [22]: df.sub(row, axis='columns')
Out[22]: 
        one       two     three
a  1.051928 -0.139606       NaN
b  0.000000  0.000000  0.000000
c  0.352192 -0.433754  1.277825
d       NaN -1.632779 -0.562782

In [23]: df.sub(row, axis=1)
Out[23]: 
        one       two     three
a  1.051928 -0.139606       NaN
b  0.000000  0.000000  0.000000
c  0.352192 -0.433754  1.277825
d       NaN -1.632779 -0.562782

In [24]: df.sub(column, axis='index')
Out[24]: 
        one  two     three
a -0.377535  0.0       NaN
b -1.569069  0.0 -1.962513
c -0.783123  0.0 -0.250933
d       NaN  0.0 -0.892516

In [25]: df.sub(column, axis=0)
Out[25]: 
        one  two     three
a -0.377535  0.0       NaN
b -1.569069  0.0 -1.962513
c -0.783123  0.0 -0.250933
d       NaN  0.0 -0.892516

Furthermore you can align a level of a MultiIndexed DataFrame with a Series.

In [26]: dfmi = df.copy()

In [27]: dfmi.index = pd.MultiIndex.from_tuples([(1, 'a'), (1, 'b'),
   ....:                                         (1, 'c'), (2, 'a')],
   ....:                                        names=['first', 'second'])
   ....: 

In [28]: dfmi.sub(column, axis=0, level='second')
Out[28]: 
                   one       two     three
first second                              
1     a      -0.377535  0.000000       NaN
      b      -1.569069  0.000000 -1.962513
      c      -0.783123  0.000000 -0.250933
2     a            NaN -1.493173 -2.385688

Series and Index also support the divmod() builtin. This function takes
the floor division and modulo operation at the same time returning a two-tuple
of the same type as the left hand side. For example:

In [29]: s = pd.Series(np.arange(10))

In [30]: s
Out[30]: 
0    0
1    1
2    2
3    3
4    4
5    5
6    6
7    7
8    8
9    9
dtype: int64

In [31]: div, rem = divmod(s, 3)

In [32]: div
Out[32]: 
0    0
1    0
2    0
3    1
4    1
5    1
6    2
7    2
8    2
9    3
dtype: int64

In [33]: rem
Out[33]: 
0    0
1    1
2    2
3    0
4    1
5    2
6    0
7    1
8    2
9    0
dtype: int64

In [34]: idx = pd.Index(np.arange(10))

In [35]: idx
Out[35]: Int64Index([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype='int64')

In [36]: div, rem = divmod(idx, 3)

In [37]: div
Out[37]: Int64Index([0, 0, 0, 1, 1, 1, 2, 2, 2, 3], dtype='int64')

In [38]: rem
Out[38]: Int64Index([0, 1, 2, 0, 1, 2, 0, 1, 2, 0], dtype='int64')

We can also do elementwise divmod():

In [39]: div, rem = divmod(s, [2, 2, 3, 3, 4, 4, 5, 5, 6, 6])

In [40]: div
Out[40]: 
0    0
1    0
2    0
3    1
4    1
5    1
6    1
7    1
8    1
9    1
dtype: int64

In [41]: rem
Out[41]: 
0    0
1    1
2    2
3    0
4    0
5    1
6    1
7    2
8    2
9    3
dtype: int64

Missing data / operations with fill values¶

In Series and DataFrame, the arithmetic functions have the option of inputting
a fill_value, namely a value to substitute when at most one of the values at
a location are missing. For example, when adding two DataFrame objects, you may
wish to treat NaN as 0 unless both DataFrames are missing that value, in which
case the result will be NaN (you can later replace NaN with some other value
using fillna if you wish).

In [42]: df
Out[42]: 
        one       two     three
a  1.394981  1.772517       NaN
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [43]: df2
Out[43]: 
        one       two     three
a  1.394981  1.772517  1.000000
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [44]: df + df2
Out[44]: 
        one       two     three
a  2.789963  3.545034       NaN
b  0.686107  3.824246 -0.100780
c  1.390491  2.956737  2.454870
d       NaN  0.558688 -1.226343

In [45]: df.add(df2, fill_value=0)
Out[45]: 
        one       two     three
a  2.789963  3.545034  1.000000
b  0.686107  3.824246 -0.100780
c  1.390491  2.956737  2.454870
d       NaN  0.558688 -1.226343

Flexible comparisons¶

Series and DataFrame have the binary comparison methods eq, ne, lt, gt,
le, and ge whose behavior is analogous to the binary
arithmetic operations described above:

In [46]: df.gt(df2)
Out[46]: 
     one    two  three
a  False  False  False
b  False  False  False
c  False  False  False
d  False  False  False

In [47]: df2.ne(df)
Out[47]: 
     one    two  three
a  False  False   True
b  False  False  False
c  False  False  False
d   True  False  False

These operations produce a pandas object of the same type as the left-hand-side
input that is of dtype bool. These boolean objects can be used in
indexing operations, see the section on Boolean indexing.

Boolean reductions¶

You can apply the reductions: empty, any(),
all(), and bool() to provide a
way to summarize a boolean result.

In [48]: (df > 0).all()
Out[48]: 
one      False
two       True
three    False
dtype: bool

In [49]: (df > 0).any()
Out[49]: 
one      True
two      True
three    True
dtype: bool

You can reduce to a final boolean value.

In [50]: (df > 0).any().any()
Out[50]: True

You can test if a pandas object is empty, via the empty property.

In [51]: df.empty
Out[51]: False

In [52]: pd.DataFrame(columns=list('ABC')).empty
Out[52]: True

To evaluate single-element pandas objects in a boolean context, use the method
bool():

In [53]: pd.Series([True]).bool()
Out[53]: True

In [54]: pd.Series([False]).bool()
Out[54]: False

In [55]: pd.DataFrame([[True]]).bool()
Out[55]: True

In [56]: pd.DataFrame([[False]]).bool()
Out[56]: False

>>> if df:
...     pass

>>> df and df2

ValueError: The truth value of an array is ambiguous. Use a.empty, a.any() or a.all().

See gotchas for a more detailed discussion.

Comparing if objects are equivalent¶

Often you may find that there is more than one way to compute the same
result. As a simple example, consider df + df and df * 2. To test
that these two computations produce the same result, given the tools
shown above, you might imagine using (df + df == df * 2).all(). But in
fact, this expression is False:

In [57]: df + df == df * 2
Out[57]: 
     one   two  three
a   True  True  False
b   True  True   True
c   True  True   True
d  False  True   True

In [58]: (df + df == df * 2).all()
Out[58]: 
one      False
two       True
three    False
dtype: bool

Notice that the boolean DataFrame df + df == df * 2 contains some False values!
This is because NaNs do not compare as equals:

In [59]: np.nan == np.nan
Out[59]: False

So, NDFrames (such as Series and DataFrames)
have an equals() method for testing equality, with NaNs in
corresponding locations treated as equal.

In [60]: (df + df).equals(df * 2)
Out[60]: True

Note that the Series or DataFrame index needs to be in the same order for
equality to be True:

In [61]: df1 = pd.DataFrame({'col': ['foo', 0, np.nan]})

In [62]: df2 = pd.DataFrame({'col': [np.nan, 0, 'foo']}, index=[2, 1, 0])

In [63]: df1.equals(df2)
Out[63]: False

In [64]: df1.equals(df2.sort_index())
Out[64]: True

Comparing array-like objects¶

You can conveniently perform element-wise comparisons when comparing a pandas
data structure with a scalar value:

In [65]: pd.Series(['foo', 'bar', 'baz']) == 'foo'
Out[65]: 
0     True
1    False
2    False
dtype: bool

In [66]: pd.Index(['foo', 'bar', 'baz']) == 'foo'
Out[66]: array([ True, False, False])

Pandas also handles element-wise comparisons between different array-like
objects of the same length:

In [67]: pd.Series(['foo', 'bar', 'baz']) == pd.Index(['foo', 'bar', 'qux'])
Out[67]: 
0     True
1     True
2    False
dtype: bool

In [68]: pd.Series(['foo', 'bar', 'baz']) == np.array(['foo', 'bar', 'qux'])
Out[68]: 
0     True
1     True
2    False
dtype: bool

Trying to compare Index or Series objects of different lengths will
raise a ValueError:

In [55]: pd.Series(['foo', 'bar', 'baz']) == pd.Series(['foo', 'bar'])
ValueError: Series lengths must match to compare

In [56]: pd.Series(['foo', 'bar', 'baz']) == pd.Series(['foo'])
ValueError: Series lengths must match to compare

Note that this is different from the NumPy behavior where a comparison can
be broadcast:

In [69]: np.array([1, 2, 3]) == np.array([2])
Out[69]: array([False,  True, False])

or it can return False if broadcasting can not be done:

In [70]: np.array([1, 2, 3]) == np.array([1, 2])
Out[70]: False

Combining overlapping data sets¶

A problem occasionally arising is the combination of two similar data sets
where values in one are preferred over the other. An example would be two data
series representing a particular economic indicator where one is considered to
be of “higher quality”. However, the lower quality series might extend further
back in history or have more complete data coverage. As such, we would like to
combine two DataFrame objects where missing values in one DataFrame are
conditionally filled with like-labeled values from the other DataFrame. The
function implementing this operation is combine_first(),
which we illustrate:

In [71]: df1 = pd.DataFrame({'A': [1., np.nan, 3., 5., np.nan],
   ....:                     'B': [np.nan, 2., 3., np.nan, 6.]})
   ....: 

In [72]: df2 = pd.DataFrame({'A': [5., 2., 4., np.nan, 3., 7.],
   ....:                     'B': [np.nan, np.nan, 3., 4., 6., 8.]})
   ....: 

In [73]: df1
Out[73]: 
     A    B
0  1.0  NaN
1  NaN  2.0
2  3.0  3.0
3  5.0  NaN
4  NaN  6.0

In [74]: df2
Out[74]: 
     A    B
0  5.0  NaN
1  2.0  NaN
2  4.0  3.0
3  NaN  4.0
4  3.0  6.0
5  7.0  8.0

In [75]: df1.combine_first(df2)
Out[75]: 
     A    B
0  1.0  NaN
1  2.0  2.0
2  3.0  3.0
3  5.0  4.0
4  3.0  6.0
5  7.0  8.0

General DataFrame combine¶

The combine_first() method above calls the more general
DataFrame.combine(). This method takes another DataFrame
and a combiner function, aligns the input DataFrame and then passes the combiner
function pairs of Series (i.e., columns whose names are the same).

So, for instance, to reproduce combine_first() as above:

In [76]: def combiner(x, y):
   ....:     return np.where(pd.isna(x), y, x)
   ....: 

Summarizing data: describe¶

There is a convenient describe() function which computes a variety of summary
statistics about a Series or the columns of a DataFrame (excluding NAs of
course):

In [93]: series = pd.Series(np.random.randn(1000))

In [94]: series[::2] = np.nan

In [95]: series.describe()
Out[95]: 
count    500.000000
mean      -0.021292
std        1.015906
min       -2.683763
25%       -0.699070
50%       -0.069718
75%        0.714483
max        3.160915
dtype: float64

In [96]: frame = pd.DataFrame(np.random.randn(1000, 5),
   ....:                      columns=['a', 'b', 'c', 'd', 'e'])
   ....: 

In [97]: frame.iloc[::2] = np.nan

In [98]: frame.describe()
Out[98]: 
                a           b           c           d           e
count  500.000000  500.000000  500.000000  500.000000  500.000000
mean     0.033387    0.030045   -0.043719   -0.051686    0.005979
std      1.017152    0.978743    1.025270    1.015988    1.006695
min     -3.000951   -2.637901   -3.303099   -3.159200   -3.188821
25%     -0.647623   -0.576449   -0.712369   -0.691338   -0.691115
50%      0.047578   -0.021499   -0.023888   -0.032652   -0.025363
75%      0.729907    0.775880    0.618896    0.670047    0.649748
max      2.740139    2.752332    3.004229    2.728702    3.240991

You can select specific percentiles to include in the output:

In [99]: series.describe(percentiles=[.05, .25, .75, .95])
Out[99]: 
count    500.000000
mean      -0.021292
std        1.015906
min       -2.683763
5%        -1.645423
25%       -0.699070
50%       -0.069718
75%        0.714483
95%        1.711409
max        3.160915
dtype: float64

By default, the median is always included.

For a non-numerical Series object, describe() will give a simple
summary of the number of unique values and most frequently occurring values:

In [100]: s = pd.Series(['a', 'a', 'b', 'b', 'a', 'a', np.nan, 'c', 'd', 'a'])

In [101]: s.describe()
Out[101]: 
count     9
unique    4
top       a
freq      5
dtype: object

Note that on a mixed-type DataFrame object, describe() will
restrict the summary to include only numerical columns or, if none are, only
categorical columns:

In [102]: frame = pd.DataFrame({'a': ['Yes', 'Yes', 'No', 'No'], 'b': range(4)})

In [103]: frame.describe()
Out[103]: 
              b
count  4.000000
mean   1.500000
std    1.290994
min    0.000000
25%    0.750000
50%    1.500000
75%    2.250000
max    3.000000

This behavior can be controlled by providing a list of types as include/exclude
arguments. The special value all can also be used:

In [104]: frame.describe(include=['object'])
Out[104]: 
          a
count     4
unique    2
top     Yes
freq      2

In [105]: frame.describe(include=['number'])
Out[105]: 
              b
count  4.000000
mean   1.500000
std    1.290994
min    0.000000
25%    0.750000
50%    1.500000
75%    2.250000
max    3.000000

In [106]: frame.describe(include='all')
Out[106]: 
          a         b
count     4  4.000000
unique    2       NaN
top     Yes       NaN
freq      2       NaN
mean    NaN  1.500000
std     NaN  1.290994
min     NaN  0.000000
25%     NaN  0.750000
50%     NaN  1.500000
75%     NaN  2.250000
max     NaN  3.000000

That feature relies on select_dtypes. Refer to
there for details about accepted inputs.

Index of min/max values¶

The idxmin() and idxmax() functions on Series
and DataFrame compute the index labels with the minimum and maximum
corresponding values:

In [107]: s1 = pd.Series(np.random.randn(5))

In [108]: s1
Out[108]: 
0    1.118076
1   -0.352051
2   -1.242883
3   -1.277155
4   -0.641184
dtype: float64

In [109]: s1.idxmin(), s1.idxmax()
Out[109]: (3, 0)

In [110]: df1 = pd.DataFrame(np.random.randn(5, 3), columns=['A', 'B', 'C'])

In [111]: df1
Out[111]: 
          A         B         C
0 -0.327863 -0.946180 -0.137570
1 -0.186235 -0.257213 -0.486567
2 -0.507027 -0.871259 -0.111110
3  2.000339 -2.430505  0.089759
4 -0.321434 -0.033695  0.096271

In [112]: df1.idxmin(axis=0)
Out[112]: 
A    2
B    3
C    1
dtype: int64

In [113]: df1.idxmax(axis=1)
Out[113]: 
0    C
1    A
2    C
3    A
4    C
dtype: object

When there are multiple rows (or columns) matching the minimum or maximum
value, idxmin() and idxmax() return the first
matching index:

In [114]: df3 = pd.DataFrame([2, 1, 1, 3, np.nan], columns=['A'], index=list('edcba'))

In [115]: df3
Out[115]: 
     A
e  2.0
d  1.0
c  1.0
b  3.0
a  NaN

In [116]: df3['A'].idxmin()
Out[116]: 'd'

Value counts (histogramming) / mode¶

The value_counts() Series method and top-level function computes a histogram
of a 1D array of values. It can also be used as a function on regular arrays:

In [117]: data = np.random.randint(0, 7, size=50)

In [118]: data
Out[118]: 
array([6, 6, 2, 3, 5, 3, 2, 5, 4, 5, 4, 3, 4, 5, 0, 2, 0, 4, 2, 0, 3, 2,
       2, 5, 6, 5, 3, 4, 6, 4, 3, 5, 6, 4, 3, 6, 2, 6, 6, 2, 3, 4, 2, 1,
       6, 2, 6, 1, 5, 4])

In [119]: s = pd.Series(data)

In [120]: s.value_counts()
Out[120]: 
6    10
2    10
4     9
5     8
3     8
0     3
1     2
dtype: int64

In [121]: pd.value_counts(data)
Out[121]: 
6    10
2    10
4     9
5     8
3     8
0     3
1     2
dtype: int64

Similarly, you can get the most frequently occurring value(s) (the mode) of the values in a Series or DataFrame:

In [122]: s5 = pd.Series([1, 1, 3, 3, 3, 5, 5, 7, 7, 7])

In [123]: s5.mode()
Out[123]: 
0    3
1    7
dtype: int64

In [124]: df5 = pd.DataFrame({"A": np.random.randint(0, 7, size=50),
   .....:                     "B": np.random.randint(-10, 15, size=50)})
   .....: 

In [125]: df5.mode()
Out[125]: 
     A   B
0  1.0  -9
1  NaN  10
2  NaN  13

Discretization and quantiling¶

Continuous values can be discretized using the cut() (bins based on values)
and qcut() (bins based on sample quantiles) functions:

In [126]: arr = np.random.randn(20)

In [127]: factor = pd.cut(arr, 4)

In [128]: factor
Out[128]: 
[(-0.251, 0.464], (-0.968, -0.251], (0.464, 1.179], (-0.251, 0.464], (-0.968, -0.251], ..., (-0.251, 0.464], (-0.968, -0.251], (-0.968, -0.251], (-0.968, -0.251], (-0.968, -0.251]]
Length: 20
Categories (4, interval[float64]): [(-0.968, -0.251] < (-0.251, 0.464] < (0.464, 1.179] <
                                    (1.179, 1.893]]

In [129]: factor = pd.cut(arr, [-5, -1, 0, 1, 5])

In [130]: factor
Out[130]: 
[(0, 1], (-1, 0], (0, 1], (0, 1], (-1, 0], ..., (-1, 0], (-1, 0], (-1, 0], (-1, 0], (-1, 0]]
Length: 20
Categories (4, interval[int64]): [(-5, -1] < (-1, 0] < (0, 1] < (1, 5]]

qcut() computes sample quantiles. For example, we could slice up some
normally distributed data into equal-size quartiles like so:

In [131]: arr = np.random.randn(30)

In [132]: factor = pd.qcut(arr, [0, .25, .5, .75, 1])

In [133]: factor
Out[133]: 
[(0.569, 1.184], (-2.278, -0.301], (-2.278, -0.301], (0.569, 1.184], (0.569, 1.184], ..., (-0.301, 0.569], (1.184, 2.346], (1.184, 2.346], (-0.301, 0.569], (-2.278, -0.301]]
Length: 30
Categories (4, interval[float64]): [(-2.278, -0.301] < (-0.301, 0.569] < (0.569, 1.184] <
                                    (1.184, 2.346]]

In [134]: pd.value_counts(factor)
Out[134]: 
(1.184, 2.346]      8
(-2.278, -0.301]    8
(0.569, 1.184]      7
(-0.301, 0.569]     7
dtype: int64

We can also pass infinite values to define the bins:

In [135]: arr = np.random.randn(20)

In [136]: factor = pd.cut(arr, [-np.inf, 0, np.inf])

In [137]: factor
Out[137]: 
[(-inf, 0.0], (0.0, inf], (0.0, inf], (-inf, 0.0], (-inf, 0.0], ..., (-inf, 0.0], (-inf, 0.0], (-inf, 0.0], (0.0, inf], (0.0, inf]]
Length: 20
Categories (2, interval[float64]): [(-inf, 0.0] < (0.0, inf]]

Tablewise function application¶

DataFrames and Series can of course just be passed into functions.
However, if the function needs to be called in a chain, consider using the pipe() method.
Compare the following

# f, g, and h are functions taking and returning ``DataFrames``
>>> f(g(h(df), arg1=1), arg2=2, arg3=3)

with the equivalent

>>> (df.pipe(h)
...    .pipe(g, arg1=1)
...    .pipe(f, arg2=2, arg3=3))

Pandas encourages the second style, which is known as method chaining.
pipe makes it easy to use your own or another library’s functions
in method chains, alongside pandas’ methods.

In the example above, the functions f, g, and h each expected the DataFrame as the first positional argument.
What if the function you wish to apply takes its data as, say, the second argument?
In this case, provide pipe with a tuple of (callable, data_keyword).
.pipe will route the DataFrame to the argument specified in the tuple.

For example, we can fit a regression using statsmodels. Their API expects a formula first and a DataFrame as the second argument, data. We pass in the function, keyword pair (sm.ols, 'data') to pipe:

In [138]: import statsmodels.formula.api as sm

In [139]: bb = pd.read_csv('data/baseball.csv', index_col='id')

In [140]: (bb.query('h > 0')
   .....:    .assign(ln_h=lambda df: np.log(df.h))
   .....:    .pipe((sm.ols, 'data'), 'hr ~ ln_h + year + g + C(lg)')
   .....:    .fit()
   .....:    .summary()
   .....:  )
   .....: 
Out[140]: 
<class 'statsmodels.iolib.summary.Summary'>
"""
                            OLS Regression Results                            
==============================================================================
Dep. Variable:                     hr   R-squared:                       0.685
Model:                            OLS   Adj. R-squared:                  0.665
Method:                 Least Squares   F-statistic:                     34.28
Date:                Sat, 09 Nov 2019   Prob (F-statistic):           3.48e-15
Time:                        19:46:29   Log-Likelihood:                -205.92
No. Observations:                  68   AIC:                             421.8
Df Residuals:                      63   BIC:                             432.9
Df Model:                           4                                         
Covariance Type:            nonrobust                                         
===============================================================================
                  coef    std err          t      P>|t|      [0.025      0.975]
-------------------------------------------------------------------------------
Intercept   -8484.7720   4664.146     -1.819      0.074   -1.78e+04     835.780
C(lg)[T.NL]    -2.2736      1.325     -1.716      0.091      -4.922       0.375
ln_h           -1.3542      0.875     -1.547      0.127      -3.103       0.395
year            4.2277      2.324      1.819      0.074      -0.417       8.872
g               0.1841      0.029      6.258      0.000       0.125       0.243
==============================================================================
Omnibus:                       10.875   Durbin-Watson:                   1.999
Prob(Omnibus):                  0.004   Jarque-Bera (JB):               17.298
Skew:                           0.537   Prob(JB):                     0.000175
Kurtosis:                       5.225   Cond. No.                     1.49e+07
==============================================================================

Warnings:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 1.49e+07. This might indicate that there are
strong multicollinearity or other numerical problems.
"""

The pipe method is inspired by unix pipes and more recently dplyr and magrittr, which
have introduced the popular (%>%) (read pipe) operator for R.
The implementation of pipe here is quite clean and feels right at home in python.
We encourage you to view the source code of pipe().

Row or column-wise function application¶

Arbitrary functions can be applied along the axes of a DataFrame
using the apply() method, which, like the descriptive
statistics methods, takes an optional axis argument:

In [141]: df.apply(np.mean)
Out[141]: 
one      0.811094
two      1.360588
three    0.187958
dtype: float64

In [142]: df.apply(np.mean, axis=1)
Out[142]: 
a    1.583749
b    0.734929
c    1.133683
d   -0.166914
dtype: float64

In [143]: df.apply(lambda x: x.max() - x.min())
Out[143]: 
one      1.051928
two      1.632779
three    1.840607
dtype: float64

In [144]: df.apply(np.cumsum)
Out[144]: 
        one       two     three
a  1.394981  1.772517       NaN
b  1.738035  3.684640 -0.050390
c  2.433281  5.163008  1.177045
d       NaN  5.442353  0.563873

In [145]: df.apply(np.exp)
Out[145]: 
        one       two     three
a  4.034899  5.885648       NaN
b  1.409244  6.767440  0.950858
c  2.004201  4.385785  3.412466
d       NaN  1.322262  0.541630

The apply() method will also dispatch on a string method name.

In [146]: df.apply('mean')
Out[146]: 
one      0.811094
two      1.360588
three    0.187958
dtype: float64

In [147]: df.apply('mean', axis=1)
Out[147]: 
a    1.583749
b    0.734929
c    1.133683
d   -0.166914
dtype: float64

The return type of the function passed to apply() affects the
type of the final output from DataFrame.apply for the default behaviour:

• If the applied function returns a Series, the final output is a DataFrame.
  The columns match the index of the Series returned by the applied function.
• If the applied function returns any other type, the final output is a Series.

This default behaviour can be overridden using the result_type, which
accepts three options: reduce, broadcast, and expand.
These will determine how list-likes return values expand (or not) to a DataFrame.

apply() combined with some cleverness can be used to answer many questions
about a data set. For example, suppose we wanted to extract the date where the
maximum value for each column occurred:

In [148]: tsdf = pd.DataFrame(np.random.randn(1000, 3), columns=['A', 'B', 'C'],
   .....:                     index=pd.date_range('1/1/2000', periods=1000))
   .....: 

In [149]: tsdf.apply(lambda x: x.idxmax())
Out[149]: 
A   2000-08-06
B   2001-01-18
C   2001-07-18
dtype: datetime64[ns]

You may also pass additional arguments and keyword arguments to the apply()
method. For instance, consider the following function you would like to apply:

def subtract_and_divide(x, sub, divide=1):
    return (x - sub) / divide

You may then apply this function as follows:

df.apply(subtract_and_divide, args=(5,), divide=3)

Another useful feature is the ability to pass Series methods to carry out some
Series operation on each column or row:

In [150]: tsdf
Out[150]: 
                   A         B         C
2000-01-01 -0.158131 -0.232466  0.321604
2000-01-02 -1.810340 -3.105758  0.433834
2000-01-03 -1.209847 -1.156793 -0.136794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08 -0.653602  0.178875  1.008298
2000-01-09  1.007996  0.462824  0.254472
2000-01-10  0.307473  0.600337  1.643950

In [151]: tsdf.apply(pd.Series.interpolate)
Out[151]: 
                   A         B         C
2000-01-01 -0.158131 -0.232466  0.321604
2000-01-02 -1.810340 -3.105758  0.433834
2000-01-03 -1.209847 -1.156793 -0.136794
2000-01-04 -1.098598 -0.889659  0.092225
2000-01-05 -0.987349 -0.622526  0.321243
2000-01-06 -0.876100 -0.355392  0.550262
2000-01-07 -0.764851 -0.088259  0.779280
2000-01-08 -0.653602  0.178875  1.008298
2000-01-09  1.007996  0.462824  0.254472
2000-01-10  0.307473  0.600337  1.643950

Finally, apply() takes an argument raw which is False by default, which
converts each row or column into a Series before applying the function. When
set to True, the passed function will instead receive an ndarray object, which
has positive performance implications if you do not need the indexing
functionality.

Aggregation API¶

The aggregation API allows one to express possibly multiple aggregation operations in a single concise way.
This API is similar across pandas objects, see groupby API, the
window functions API, and the resample API.
The entry point for aggregation is DataFrame.aggregate(), or the alias
DataFrame.agg().

We will use a similar starting frame from above:

In [152]: tsdf = pd.DataFrame(np.random.randn(10, 3), columns=['A', 'B', 'C'],
   .....:                     index=pd.date_range('1/1/2000', periods=10))
   .....: 

In [153]: tsdf.iloc[3:7] = np.nan

In [154]: tsdf
Out[154]: 
                   A         B         C
2000-01-01  1.257606  1.004194  0.167574
2000-01-02 -0.749892  0.288112 -0.757304
2000-01-03 -0.207550 -0.298599  0.116018
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.814347 -0.257623  0.869226
2000-01-09 -0.250663 -1.206601  0.896839
2000-01-10  2.169758 -1.333363  0.283157

Using a single function is equivalent to apply(). You can also
pass named methods as strings. These will return a Series of the aggregated
output:

In [155]: tsdf.agg(np.sum)
Out[155]: 
A    3.033606
B   -1.803879
C    1.575510
dtype: float64

In [156]: tsdf.agg('sum')
Out[156]: 
A    3.033606
B   -1.803879
C    1.575510
dtype: float64

# these are equivalent to a ``.sum()`` because we are aggregating
# on a single function
In [157]: tsdf.sum()
Out[157]: 
A    3.033606
B   -1.803879
C    1.575510
dtype: float64

Single aggregations on a Series this will return a scalar value:

In [158]: tsdf.A.agg('sum')
Out[158]: 3.033606102414146

In [159]: tsdf.agg(['sum'])
Out[159]: 
            A         B        C
sum  3.033606 -1.803879  1.57551

In [160]: tsdf.agg(['sum', 'mean'])
Out[160]: 
             A         B         C
sum   3.033606 -1.803879  1.575510
mean  0.505601 -0.300647  0.262585

In [161]: tsdf.A.agg(['sum', 'mean'])
Out[161]: 
sum     3.033606
mean    0.505601
Name: A, dtype: float64

In [162]: tsdf.A.agg(['sum', lambda x: x.mean()])
Out[162]: 
sum         3.033606
<lambda>    0.505601
Name: A, dtype: float64

In [163]: def mymean(x):
   .....:     return x.mean()
   .....: 

In [164]: tsdf.A.agg(['sum', mymean])
Out[164]: 
sum       3.033606
mymean    0.505601
Name: A, dtype: float64

In [165]: tsdf.agg({'A': 'mean', 'B': 'sum'})
Out[165]: 
A    0.505601
B   -1.803879
dtype: float64

In [166]: tsdf.agg({'A': ['mean', 'min'], 'B': 'sum'})
Out[166]: 
             A         B
mean  0.505601       NaN
min  -0.749892       NaN
sum        NaN -1.803879

In [167]: mdf = pd.DataFrame({'A': [1, 2, 3],
   .....:                     'B': [1., 2., 3.],
   .....:                     'C': ['foo', 'bar', 'baz'],
   .....:                     'D': pd.date_range('20130101', periods=3)})
   .....: 

In [168]: mdf.dtypes
Out[168]: 
A             int64
B           float64
C            object
D    datetime64[ns]
dtype: object

In [169]: mdf.agg(['min', 'sum'])
Out[169]: 
     A    B          C          D
min  1  1.0        bar 2013-01-01
sum  6  6.0  foobarbaz        NaT

In [170]: from functools import partial

In [171]: q_25 = partial(pd.Series.quantile, q=0.25)

In [172]: q_25.__name__ = '25%'

In [173]: q_75 = partial(pd.Series.quantile, q=0.75)

In [174]: q_75.__name__ = '75%'

In [175]: tsdf.agg(['count', 'mean', 'std', 'min', q_25, 'median', q_75, 'max'])
Out[175]: 
               A         B         C
count   6.000000  6.000000  6.000000
mean    0.505601 -0.300647  0.262585
std     1.103362  0.887508  0.606860
min    -0.749892 -1.333363 -0.757304
25%    -0.239885 -0.979600  0.128907
median  0.303398 -0.278111  0.225365
75%     1.146791  0.151678  0.722709
max     2.169758  1.004194  0.896839

Transform API¶

The transform() method returns an object that is indexed the same (same size)
as the original. This API allows you to provide multiple operations at the same
time rather than one-by-one. Its API is quite similar to the .agg API.

We create a frame similar to the one used in the above sections.

In [176]: tsdf = pd.DataFrame(np.random.randn(10, 3), columns=['A', 'B', 'C'],
   .....:                     index=pd.date_range('1/1/2000', periods=10))
   .....: 

In [177]: tsdf.iloc[3:7] = np.nan

In [178]: tsdf
Out[178]: 
                   A         B         C
2000-01-01 -0.428759 -0.864890 -0.675341
2000-01-02 -0.168731  1.338144 -1.279321
2000-01-03 -1.621034  0.438107  0.903794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374 -1.240447 -0.201052
2000-01-09 -0.157795  0.791197 -1.144209
2000-01-10 -0.030876  0.371900  0.061932

Transform the entire frame. .transform() allows input functions as: a NumPy function, a string
function name or a user defined function.

In [179]: tsdf.transform(np.abs)
Out[179]: 
                   A         B         C
2000-01-01  0.428759  0.864890  0.675341
2000-01-02  0.168731  1.338144  1.279321
2000-01-03  1.621034  0.438107  0.903794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374  1.240447  0.201052
2000-01-09  0.157795  0.791197  1.144209
2000-01-10  0.030876  0.371900  0.061932

In [180]: tsdf.transform('abs')
Out[180]: 
                   A         B         C
2000-01-01  0.428759  0.864890  0.675341
2000-01-02  0.168731  1.338144  1.279321
2000-01-03  1.621034  0.438107  0.903794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374  1.240447  0.201052
2000-01-09  0.157795  0.791197  1.144209
2000-01-10  0.030876  0.371900  0.061932

In [181]: tsdf.transform(lambda x: x.abs())
Out[181]: 
                   A         B         C
2000-01-01  0.428759  0.864890  0.675341
2000-01-02  0.168731  1.338144  1.279321
2000-01-03  1.621034  0.438107  0.903794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374  1.240447  0.201052
2000-01-09  0.157795  0.791197  1.144209
2000-01-10  0.030876  0.371900  0.061932

Here transform() received a single function; this is equivalent to a ufunc application.

In [182]: np.abs(tsdf)
Out[182]: 
                   A         B         C
2000-01-01  0.428759  0.864890  0.675341
2000-01-02  0.168731  1.338144  1.279321
2000-01-03  1.621034  0.438107  0.903794
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374  1.240447  0.201052
2000-01-09  0.157795  0.791197  1.144209
2000-01-10  0.030876  0.371900  0.061932

Passing a single function to .transform() with a Series will yield a single Series in return.

In [183]: tsdf.A.transform(np.abs)
Out[183]: 
2000-01-01    0.428759
2000-01-02    0.168731
2000-01-03    1.621034
2000-01-04         NaN
2000-01-05         NaN
2000-01-06         NaN
2000-01-07         NaN
2000-01-08    0.254374
2000-01-09    0.157795
2000-01-10    0.030876
Freq: D, Name: A, dtype: float64

In [184]: tsdf.transform([np.abs, lambda x: x + 1])
Out[184]: 
                   A                   B                   C          
            absolute  <lambda>  absolute  <lambda>  absolute  <lambda>
2000-01-01  0.428759  0.571241  0.864890  0.135110  0.675341  0.324659
2000-01-02  0.168731  0.831269  1.338144  2.338144  1.279321 -0.279321
2000-01-03  1.621034 -0.621034  0.438107  1.438107  0.903794  1.903794
2000-01-04       NaN       NaN       NaN       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN       NaN       NaN       NaN
2000-01-08  0.254374  1.254374  1.240447 -0.240447  0.201052  0.798948
2000-01-09  0.157795  0.842205  0.791197  1.791197  1.144209 -0.144209
2000-01-10  0.030876  0.969124  0.371900  1.371900  0.061932  1.061932

In [185]: tsdf.A.transform([np.abs, lambda x: x + 1])
Out[185]: 
            absolute  <lambda>
2000-01-01  0.428759  0.571241
2000-01-02  0.168731  0.831269
2000-01-03  1.621034 -0.621034
2000-01-04       NaN       NaN
2000-01-05       NaN       NaN
2000-01-06       NaN       NaN
2000-01-07       NaN       NaN
2000-01-08  0.254374  1.254374
2000-01-09  0.157795  0.842205
2000-01-10  0.030876  0.969124

In [186]: tsdf.transform({'A': np.abs, 'B': lambda x: x + 1})
Out[186]: 
                   A         B
2000-01-01  0.428759  0.135110
2000-01-02  0.168731  2.338144
2000-01-03  1.621034  1.438107
2000-01-04       NaN       NaN
2000-01-05       NaN       NaN
2000-01-06       NaN       NaN
2000-01-07       NaN       NaN
2000-01-08  0.254374 -0.240447
2000-01-09  0.157795  1.791197
2000-01-10  0.030876  1.371900

In [187]: tsdf.transform({'A': np.abs, 'B': [lambda x: x + 1, 'sqrt']})
Out[187]: 
                   A         B          
            absolute  <lambda>      sqrt
2000-01-01  0.428759  0.135110       NaN
2000-01-02  0.168731  2.338144  1.156782
2000-01-03  1.621034  1.438107  0.661897
2000-01-04       NaN       NaN       NaN
2000-01-05       NaN       NaN       NaN
2000-01-06       NaN       NaN       NaN
2000-01-07       NaN       NaN       NaN
2000-01-08  0.254374 -0.240447       NaN
2000-01-09  0.157795  1.791197  0.889493
2000-01-10  0.030876  1.371900  0.609836

Applying elementwise functions¶

Since not all functions can be vectorized (accept NumPy arrays and return
another array or value), the methods applymap() on DataFrame
and analogously map() on Series accept any Python function taking
a single value and returning a single value. For example:

In [188]: df4
Out[188]: 
        one       two     three
a  1.394981  1.772517       NaN
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [189]: def f(x):
   .....:     return len(str(x))
   .....: 

In [190]: df4['one'].map(f)
Out[190]: 
a    18
b    19
c    18
d     3
Name: one, dtype: int64

In [191]: df4.applymap(f)
Out[191]: 
   one  two  three
a   18   17      3
b   19   18     20
c   18   18     16
d    3   19     19

Series.map() has an additional feature; it can be used to easily
“link” or “map” values defined by a secondary series. This is closely related
to merging/joining functionality:

In [192]: s = pd.Series(['six', 'seven', 'six', 'seven', 'six'],
   .....:               index=['a', 'b', 'c', 'd', 'e'])
   .....: 

In [193]: t = pd.Series({'six': 6., 'seven': 7.})

In [194]: s
Out[194]: 
a      six
b    seven
c      six
d    seven
e      six
dtype: object

In [195]: s.map(t)
Out[195]: 
a    6.0
b    7.0
c    6.0
d    7.0
e    6.0
dtype: float64

Reindexing to align with another object¶

You may wish to take an object and reindex its axes to be labeled the same as
another object. While the syntax for this is straightforward albeit verbose, it
is a common enough operation that the reindex_like() method is
available to make this simpler:

In [207]: df2
Out[207]: 
        one       two
a  1.394981  1.772517
b  0.343054  1.912123
c  0.695246  1.478369

In [208]: df3
Out[208]: 
        one       two
a  0.583888  0.051514
b -0.468040  0.191120
c -0.115848 -0.242634

In [209]: df.reindex_like(df2)
Out[209]: 
        one       two
a  1.394981  1.772517
b  0.343054  1.912123
c  0.695246  1.478369

Aligning objects with each other with align¶

The align() method is the fastest way to simultaneously align two objects. It
supports a join argument (related to joining and merging):

• join='outer': take the union of the indexes (default)
• join='left': use the calling object’s index
• join='right': use the passed object’s index
• join='inner': intersect the indexes

It returns a tuple with both of the reindexed Series:

In [210]: s = pd.Series(np.random.randn(5), index=['a', 'b', 'c', 'd', 'e'])

In [211]: s1 = s[:4]

In [212]: s2 = s[1:]

In [213]: s1.align(s2)
Out[213]: 
(a   -0.186646
 b   -1.692424
 c   -0.303893
 d   -1.425662
 e         NaN
 dtype: float64, a         NaN
 b   -1.692424
 c   -0.303893
 d   -1.425662
 e    1.114285
 dtype: float64)

In [214]: s1.align(s2, join='inner')
Out[214]: 
(b   -1.692424
 c   -0.303893
 d   -1.425662
 dtype: float64, b   -1.692424
 c   -0.303893
 d   -1.425662
 dtype: float64)

In [215]: s1.align(s2, join='left')
Out[215]: 
(a   -0.186646
 b   -1.692424
 c   -0.303893
 d   -1.425662
 dtype: float64, a         NaN
 b   -1.692424
 c   -0.303893
 d   -1.425662
 dtype: float64)

For DataFrames, the join method will be applied to both the index and the
columns by default:

In [216]: df.align(df2, join='inner')
Out[216]: 
(        one       two
 a  1.394981  1.772517
 b  0.343054  1.912123
 c  0.695246  1.478369,         one       two
 a  1.394981  1.772517
 b  0.343054  1.912123
 c  0.695246  1.478369)

You can also pass an axis option to only align on the specified axis:

In [217]: df.align(df2, join='inner', axis=0)
Out[217]: 
(        one       two     three
 a  1.394981  1.772517       NaN
 b  0.343054  1.912123 -0.050390
 c  0.695246  1.478369  1.227435,         one       two
 a  1.394981  1.772517
 b  0.343054  1.912123
 c  0.695246  1.478369)

If you pass a Series to DataFrame.align(), you can choose to align both
objects either on the DataFrame’s index or columns using the axis argument:

In [218]: df.align(df2.iloc[0], axis=1)
Out[218]: 
(        one     three       two
 a  1.394981       NaN  1.772517
 b  0.343054 -0.050390  1.912123
 c  0.695246  1.227435  1.478369
 d       NaN -0.613172  0.279344, one      1.394981
 three         NaN
 two      1.772517
 Name: a, dtype: float64)

Filling while reindexing¶

reindex() takes an optional parameter method which is a
filling method chosen from the following table:

We illustrate these fill methods on a simple Series:

In [219]: rng = pd.date_range('1/3/2000', periods=8)

In [220]: ts = pd.Series(np.random.randn(8), index=rng)

In [221]: ts2 = ts[[0, 3, 6]]

In [222]: ts
Out[222]: 
2000-01-03    0.183051
2000-01-04    0.400528
2000-01-05   -0.015083
2000-01-06    2.395489
2000-01-07    1.414806
2000-01-08    0.118428
2000-01-09    0.733639
2000-01-10   -0.936077
Freq: D, dtype: float64

In [223]: ts2
Out[223]: 
2000-01-03    0.183051
2000-01-06    2.395489
2000-01-09    0.733639
dtype: float64

In [224]: ts2.reindex(ts.index)
Out[224]: 
2000-01-03    0.183051
2000-01-04         NaN
2000-01-05         NaN
2000-01-06    2.395489
2000-01-07         NaN
2000-01-08         NaN
2000-01-09    0.733639
2000-01-10         NaN
Freq: D, dtype: float64

In [225]: ts2.reindex(ts.index, method='ffill')
Out[225]: 
2000-01-03    0.183051
2000-01-04    0.183051
2000-01-05    0.183051
2000-01-06    2.395489
2000-01-07    2.395489
2000-01-08    2.395489
2000-01-09    0.733639
2000-01-10    0.733639
Freq: D, dtype: float64

In [226]: ts2.reindex(ts.index, method='bfill')
Out[226]: 
2000-01-03    0.183051
2000-01-04    2.395489
2000-01-05    2.395489
2000-01-06    2.395489
2000-01-07    0.733639
2000-01-08    0.733639
2000-01-09    0.733639
2000-01-10         NaN
Freq: D, dtype: float64

In [227]: ts2.reindex(ts.index, method='nearest')
Out[227]: 
2000-01-03    0.183051
2000-01-04    0.183051
2000-01-05    2.395489
2000-01-06    2.395489
2000-01-07    2.395489
2000-01-08    0.733639
2000-01-09    0.733639
2000-01-10    0.733639
Freq: D, dtype: float64

These methods require that the indexes are ordered increasing or
decreasing.

Note that the same result could have been achieved using
fillna (except for method='nearest') or
interpolate:

In [228]: ts2.reindex(ts.index).fillna(method='ffill')
Out[228]: 
2000-01-03    0.183051
2000-01-04    0.183051
2000-01-05    0.183051
2000-01-06    2.395489
2000-01-07    2.395489
2000-01-08    2.395489
2000-01-09    0.733639
2000-01-10    0.733639
Freq: D, dtype: float64

reindex() will raise a ValueError if the index is not monotonically
increasing or decreasing. fillna() and interpolate()
will not perform any checks on the order of the index.

Limits on filling while reindexing¶

The limit and tolerance arguments provide additional control over
filling while reindexing. Limit specifies the maximum count of consecutive
matches:

In [229]: ts2.reindex(ts.index, method='ffill', limit=1)
Out[229]: 
2000-01-03    0.183051
2000-01-04    0.183051
2000-01-05         NaN
2000-01-06    2.395489
2000-01-07    2.395489
2000-01-08         NaN
2000-01-09    0.733639
2000-01-10    0.733639
Freq: D, dtype: float64

In contrast, tolerance specifies the maximum distance between the index and
indexer values:

In [230]: ts2.reindex(ts.index, method='ffill', tolerance='1 day')
Out[230]: 
2000-01-03    0.183051
2000-01-04    0.183051
2000-01-05         NaN
2000-01-06    2.395489
2000-01-07    2.395489
2000-01-08         NaN
2000-01-09    0.733639
2000-01-10    0.733639
Freq: D, dtype: float64

Notice that when used on a DatetimeIndex, TimedeltaIndex or
PeriodIndex, tolerance will coerced into a Timedelta if possible.
This allows you to specify tolerance with appropriate strings.

Dropping labels from an axis¶

A method closely related to reindex is the drop() function.
It removes a set of labels from an axis:

In [231]: df
Out[231]: 
        one       two     three
a  1.394981  1.772517       NaN
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [232]: df.drop(['a', 'd'], axis=0)
Out[232]: 
        one       two     three
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435

In [233]: df.drop(['one'], axis=1)
Out[233]: 
        two     three
a  1.772517       NaN
b  1.912123 -0.050390
c  1.478369  1.227435
d  0.279344 -0.613172

Note that the following also works, but is a bit less obvious / clean:

In [234]: df.reindex(df.index.difference(['a', 'd']))
Out[234]: 
        one       two     three
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435

Renaming / mapping labels¶

The rename() method allows you to relabel an axis based on some
mapping (a dict or Series) or an arbitrary function.

In [235]: s
Out[235]: 
a   -0.186646
b   -1.692424
c   -0.303893
d   -1.425662
e    1.114285
dtype: float64

In [236]: s.rename(str.upper)
Out[236]: 
A   -0.186646
B   -1.692424
C   -0.303893
D   -1.425662
E    1.114285
dtype: float64

If you pass a function, it must return a value when called with any of the
labels (and must produce a set of unique values). A dict or
Series can also be used:

In [237]: df.rename(columns={'one': 'foo', 'two': 'bar'},
   .....:           index={'a': 'apple', 'b': 'banana', 'd': 'durian'})
   .....: 
Out[237]: 
             foo       bar     three
apple   1.394981  1.772517       NaN
banana  0.343054  1.912123 -0.050390
c       0.695246  1.478369  1.227435
durian       NaN  0.279344 -0.613172

If the mapping doesn’t include a column/index label, it isn’t renamed. Note that
extra labels in the mapping don’t throw an error.

DataFrame.rename() also supports an “axis-style” calling convention, where
you specify a single mapper and the axis to apply that mapping to.

In [238]: df.rename({'one': 'foo', 'two': 'bar'}, axis='columns')
Out[238]: 
        foo       bar     three
a  1.394981  1.772517       NaN
b  0.343054  1.912123 -0.050390
c  0.695246  1.478369  1.227435
d       NaN  0.279344 -0.613172

In [239]: df.rename({'a': 'apple', 'b': 'banana', 'd': 'durian'}, axis='index')
Out[239]: 
             one       two     three
apple   1.394981  1.772517       NaN
banana  0.343054  1.912123 -0.050390
c       0.695246  1.478369  1.227435
durian       NaN  0.279344 -0.613172

The rename() method also provides an inplace named
parameter that is by default False and copies the underlying data. Pass
inplace=True to rename the data in place.

Finally, rename() also accepts a scalar or list-like
for altering the Series.name attribute.

In [240]: s.rename("scalar-name")
Out[240]: 
a   -0.186646
b   -1.692424
c   -0.303893
d   -1.425662
e    1.114285
Name: scalar-name, dtype: float64

The methods rename_axis() and rename_axis()
allow specific names of a MultiIndex to be changed (as opposed to the
labels).

In [241]: df = pd.DataFrame({'x': [1, 2, 3, 4, 5, 6],
   .....:                    'y': [10, 20, 30, 40, 50, 60]},
   .....:                   index=pd.MultiIndex.from_product([['a', 'b', 'c'], [1, 2]],
   .....:                   names=['let', 'num']))
   .....: 

In [242]: df
Out[242]: 
         x   y
let num       
a   1    1  10
    2    2  20
b   1    3  30
    2    4  40
c   1    5  50
    2    6  60

In [243]: df.rename_axis(index={'let': 'abc'})
Out[243]: 
         x   y
abc num       
a   1    1  10
    2    2  20
b   1    3  30
    2    4  40
c   1    5  50
    2    6  60

In [244]: df.rename_axis(index=str.upper)
Out[244]: 
         x   y
LET NUM       
a   1    1  10
    2    2  20
b   1    3  30
    2    4  40
c   1    5  50
    2    6  60

items¶

Consistent with the dict-like interface, items() iterates
through key-value pairs:

• Series: (index, scalar value) pairs
• DataFrame: (column, Series) pairs

For example:

In [250]: for label, ser in df.items():
   .....:     print(label)
   .....:     print(ser)
   .....: 
a
0    1
1    2
2    3
Name: a, dtype: int64
b
0    a
1    b
2    c
Name: b, dtype: object

iterrows¶

iterrows() allows you to iterate through the rows of a
DataFrame as Series objects. It returns an iterator yielding each
index value along with a Series containing the data in each row:

In [251]: for row_index, row in df.iterrows():
   .....:     print(row_index, row, sep='n')
   .....: 
0
a    1
b    a
Name: 0, dtype: object
1
a    2
b    b
Name: 1, dtype: object
2
a    3
b    c
Name: 2, dtype: object

In [252]: df_orig = pd.DataFrame([[1, 1.5]], columns=['int', 'float'])

In [253]: df_orig.dtypes
Out[253]: 
int        int64
float    float64
dtype: object

In [254]: row = next(df_orig.iterrows())[1]

In [255]: row
Out[255]: 
int      1.0
float    1.5
Name: 0, dtype: float64

In [256]: row['int'].dtype
Out[256]: dtype('float64')

In [257]: df_orig['int'].dtype
Out[257]: dtype('int64')

For instance, a contrived way to transpose the DataFrame would be:

In [258]: df2 = pd.DataFrame({'x': [1, 2, 3], 'y': [4, 5, 6]})

In [259]: print(df2)
   x  y
0  1  4
1  2  5
2  3  6

In [260]: print(df2.T)
  0  1  2
x  1  2  3
y  4  5  6

In [261]: df2_t = pd.DataFrame({idx: values for idx, values in df2.iterrows()})

In [262]: print(df2_t)
   0  1  2
x  1  2  3
y  4  5  6

itertuples¶

The itertuples() method will return an iterator
yielding a namedtuple for each row in the DataFrame. The first element
of the tuple will be the row’s corresponding index value, while the
remaining values are the row values.

For instance:

In [263]: for row in df.itertuples():
   .....:     print(row)
   .....: 
Pandas(Index=0, a=1, b='a')
Pandas(Index=1, a=2, b='b')
Pandas(Index=2, a=3, b='c')

This method does not convert the row to a Series object; it merely
returns the values inside a namedtuple. Therefore,
itertuples() preserves the data type of the values
and is generally faster as iterrows().

By index¶

The Series.sort_index() and DataFrame.sort_index() methods are
used to sort a pandas object by its index levels.

In [291]: df = pd.DataFrame({
   .....:     'one': pd.Series(np.random.randn(3), index=['a', 'b', 'c']),
   .....:     'two': pd.Series(np.random.randn(4), index=['a', 'b', 'c', 'd']),
   .....:     'three': pd.Series(np.random.randn(3), index=['b', 'c', 'd'])})
   .....: 

In [292]: unsorted_df = df.reindex(index=['a', 'd', 'c', 'b'],
   .....:                          columns=['three', 'two', 'one'])
   .....: 

In [293]: unsorted_df
Out[293]: 
      three       two       one
a       NaN -1.152244  0.562973
d -0.252916 -0.109597       NaN
c  1.273388 -0.167123  0.640382
b -0.098217  0.009797 -1.299504

# DataFrame
In [294]: unsorted_df.sort_index()
Out[294]: 
      three       two       one
a       NaN -1.152244  0.562973
b -0.098217  0.009797 -1.299504
c  1.273388 -0.167123  0.640382
d -0.252916 -0.109597       NaN

In [295]: unsorted_df.sort_index(ascending=False)
Out[295]: 
      three       two       one
d -0.252916 -0.109597       NaN
c  1.273388 -0.167123  0.640382
b -0.098217  0.009797 -1.299504
a       NaN -1.152244  0.562973

In [296]: unsorted_df.sort_index(axis=1)
Out[296]: 
        one     three       two
a  0.562973       NaN -1.152244
d       NaN -0.252916 -0.109597
c  0.640382  1.273388 -0.167123
b -1.299504 -0.098217  0.009797

# Series
In [297]: unsorted_df['three'].sort_index()
Out[297]: 
a         NaN
b   -0.098217
c    1.273388
d   -0.252916
Name: three, dtype: float64

By values¶

The Series.sort_values() method is used to sort a Series by its values. The
DataFrame.sort_values() method is used to sort a DataFrame by its column or row values.
The optional by parameter to DataFrame.sort_values() may used to specify one or more columns
to use to determine the sorted order.

In [298]: df1 = pd.DataFrame({'one': [2, 1, 1, 1],
   .....:                     'two': [1, 3, 2, 4],
   .....:                     'three': [5, 4, 3, 2]})
   .....: 

In [299]: df1.sort_values(by='two')
Out[299]: 
   one  two  three
0    2    1      5
2    1    2      3
1    1    3      4
3    1    4      2

The by parameter can take a list of column names, e.g.:

In [300]: df1[['one', 'two', 'three']].sort_values(by=['one', 'two'])
Out[300]: 
   one  two  three
2    1    2      3
1    1    3      4
3    1    4      2
0    2    1      5

These methods have special treatment of NA values via the na_position
argument:

In [301]: s[2] = np.nan

In [302]: s.sort_values()
Out[302]: 
0       A
3    Aaba
1       B
4    Baca
6    CABA
8     cat
7     dog
2     NaN
5     NaN
dtype: object

In [303]: s.sort_values(na_position='first')
Out[303]: 
2     NaN
5     NaN
0       A
3    Aaba
1       B
4    Baca
6    CABA
8     cat
7     dog
dtype: object

By indexes and values¶

Strings passed as the by parameter to DataFrame.sort_values() may
refer to either columns or index level names.

# Build MultiIndex
In [304]: idx = pd.MultiIndex.from_tuples([('a', 1), ('a', 2), ('a', 2),
   .....:                                 ('b', 2), ('b', 1), ('b', 1)])
   .....: 

In [305]: idx.names = ['first', 'second']

# Build DataFrame
In [306]: df_multi = pd.DataFrame({'A': np.arange(6, 0, -1)},
   .....:                         index=idx)
   .....: 

In [307]: df_multi
Out[307]: 
              A
first second   
a     1       6
      2       5
      2       4
b     2       3
      1       2
      1       1

Sort by ‘second’ (index) and ‘A’ (column)

In [308]: df_multi.sort_values(by=['second', 'A'])
Out[308]: 
              A
first second   
b     1       1
      1       2
a     1       6
b     2       3
a     2       4
      2       5

searchsorted¶

Series has the searchsorted() method, which works similarly to
numpy.ndarray.searchsorted().

In [309]: ser = pd.Series([1, 2, 3])

In [310]: ser.searchsorted([0, 3])
Out[310]: array([0, 2])

In [311]: ser.searchsorted([0, 4])
Out[311]: array([0, 3])

In [312]: ser.searchsorted([1, 3], side='right')
Out[312]: array([1, 3])

In [313]: ser.searchsorted([1, 3], side='left')
Out[313]: array([0, 2])

In [314]: ser = pd.Series([3, 1, 2])

In [315]: ser.searchsorted([0, 3], sorter=np.argsort(ser))
Out[315]: array([0, 2])

smallest / largest values¶

Series has the nsmallest() and nlargest() methods which return the
smallest or largest (n) values. For a large Series this can be much
faster than sorting the entire Series and calling head(n) on the result.

In [316]: s = pd.Series(np.random.permutation(10))

In [317]: s
Out[317]: 
0    2
1    0
2    3
3    7
4    1
5    5
6    9
7    6
8    8
9    4
dtype: int64

In [318]: s.sort_values()
Out[318]: 
1    0
4    1
0    2
2    3
9    4
5    5
7    6
3    7
8    8
6    9
dtype: int64

In [319]: s.nsmallest(3)
Out[319]: 
1    0
4    1
0    2
dtype: int64

In [320]: s.nlargest(3)
Out[320]: 
6    9
8    8
3    7
dtype: int64

DataFrame also has the nlargest and nsmallest methods.

In [321]: df = pd.DataFrame({'a': [-2, -1, 1, 10, 8, 11, -1],
   .....:                    'b': list('abdceff'),
   .....:                    'c': [1.0, 2.0, 4.0, 3.2, np.nan, 3.0, 4.0]})
   .....: 

In [322]: df.nlargest(3, 'a')
Out[322]: 
    a  b    c
5  11  f  3.0
3  10  c  3.2
4   8  e  NaN

In [323]: df.nlargest(5, ['a', 'c'])
Out[323]: 
    a  b    c
5  11  f  3.0
3  10  c  3.2
4   8  e  NaN
2   1  d  4.0
6  -1  f  4.0

In [324]: df.nsmallest(3, 'a')
Out[324]: 
   a  b    c
0 -2  a  1.0
1 -1  b  2.0
6 -1  f  4.0

In [325]: df.nsmallest(5, ['a', 'c'])
Out[325]: 
   a  b    c
0 -2  a  1.0
1 -1  b  2.0
6 -1  f  4.0
2  1  d  4.0
4  8  e  NaN

Sorting by a MultiIndex column¶

You must be explicit about sorting when the column is a MultiIndex, and fully specify
all levels to by.

In [326]: df1.columns = pd.MultiIndex.from_tuples([('a', 'one'),
   .....:                                          ('a', 'two'),
   .....:                                          ('b', 'three')])
   .....: 

In [327]: df1.sort_values(by=('a', 'two'))
Out[327]: 
    a         b
  one two three
0   2   1     5
2   1   2     3
1   1   3     4
3   1   4     2

defaults¶

By default integer types are int64 and float types are float64,
regardless of platform (32-bit or 64-bit).
The following will all result in int64 dtypes.

In [341]: pd.DataFrame([1, 2], columns=['a']).dtypes
Out[341]: 
a    int64
dtype: object

In [342]: pd.DataFrame({'a': [1, 2]}).dtypes
Out[342]: 
a    int64
dtype: object

In [343]: pd.DataFrame({'a': 1}, index=list(range(2))).dtypes
Out[343]: 
a    int64
dtype: object

Note that Numpy will choose platform-dependent types when creating arrays.
The following WILL result in int32 on 32-bit platform.

In [344]: frame = pd.DataFrame(np.array([1, 2]))

upcasting¶

Types can potentially be upcasted when combined with other types, meaning they are promoted
from the current type (e.g. int to float).

In [345]: df3 = df1.reindex_like(df2).fillna(value=0.0) + df2

In [346]: df3
Out[346]: 
          A         B      C
0  1.047606  0.256090    0.0
1  3.497968  1.426469    0.0
2 -0.150862 -0.416203  255.0
3  0.724370  1.139976    0.0
4 -1.203098 -1.193477    0.0
5  1.346426  0.096706    0.0
6 -0.052599 -1.956850    1.0
7 -0.756495 -0.714337    0.0

In [347]: df3.dtypes
Out[347]: 
A    float32
B    float64
C    float64
dtype: object

DataFrame.to_numpy() will return the lower-common-denominator of the dtypes, meaning
the dtype that can accommodate ALL of the types in the resulting homogeneous dtyped NumPy array. This can
force some upcasting.

In [348]: df3.to_numpy().dtype
Out[348]: dtype('float64')

astype¶

You can use the astype() method to explicitly convert dtypes from one to another. These will by default return a copy,
even if the dtype was unchanged (pass copy=False to change this behavior). In addition, they will raise an
exception if the astype operation is invalid.

Upcasting is always according to the numpy rules. If two different dtypes are involved in an operation,
then the more general one will be used as the result of the operation.

In [349]: df3
Out[349]: 
          A         B      C
0  1.047606  0.256090    0.0
1  3.497968  1.426469    0.0
2 -0.150862 -0.416203  255.0
3  0.724370  1.139976    0.0
4 -1.203098 -1.193477    0.0
5  1.346426  0.096706    0.0
6 -0.052599 -1.956850    1.0
7 -0.756495 -0.714337    0.0

In [350]: df3.dtypes
Out[350]: 
A    float32
B    float64
C    float64
dtype: object

# conversion of dtypes
In [351]: df3.astype('float32').dtypes
Out[351]: 
A    float32
B    float32
C    float32
dtype: object

Convert a subset of columns to a specified type using astype().

In [352]: dft = pd.DataFrame({'a': [1, 2, 3], 'b': [4, 5, 6], 'c': [7, 8, 9]})

In [353]: dft[['a', 'b']] = dft[['a', 'b']].astype(np.uint8)

In [354]: dft
Out[354]: 
   a  b  c
0  1  4  7
1  2  5  8
2  3  6  9

In [355]: dft.dtypes
Out[355]: 
a    uint8
b    uint8
c    int64
dtype: object

Convert certain columns to a specific dtype by passing a dict to astype().

In [356]: dft1 = pd.DataFrame({'a': [1, 0, 1], 'b': [4, 5, 6], 'c': [7, 8, 9]})

In [357]: dft1 = dft1.astype({'a': np.bool, 'c': np.float64})

In [358]: dft1
Out[358]: 
       a  b    c
0   True  4  7.0
1  False  5  8.0
2   True  6  9.0

In [359]: dft1.dtypes
Out[359]: 
a       bool
b      int64
c    float64
dtype: object

In [360]: dft = pd.DataFrame({'a': [1, 2, 3], 'b': [4, 5, 6], 'c': [7, 8, 9]})

In [361]: dft.loc[:, ['a', 'b']].astype(np.uint8).dtypes
Out[361]: 
a    uint8
b    uint8
dtype: object

In [362]: dft.loc[:, ['a', 'b']] = dft.loc[:, ['a', 'b']].astype(np.uint8)

In [363]: dft.dtypes
Out[363]: 
a    int64
b    int64
c    int64
dtype: object

object conversion¶

pandas offers various functions to try to force conversion of types from the object dtype to other types.
In cases where the data is already of the correct type, but stored in an object array, the
DataFrame.infer_objects() and Series.infer_objects() methods can be used to soft convert
to the correct type.

In [364]: import datetime

In [365]: df = pd.DataFrame([[1, 2],
   .....:                    ['a', 'b'],
   .....:                    [datetime.datetime(2016, 3, 2),
   .....:                     datetime.datetime(2016, 3, 2)]])
   .....: 

In [366]: df = df.T

In [367]: df
Out[367]: 
   0  1          2
0  1  a 2016-03-02
1  2  b 2016-03-02

In [368]: df.dtypes
Out[368]: 
0            object
1            object
2    datetime64[ns]
dtype: object

Because the data was transposed the original inference stored all columns as object, which
infer_objects will correct.

In [369]: df.infer_objects().dtypes
Out[369]: 
0             int64
1            object
2    datetime64[ns]
dtype: object

The following functions are available for one dimensional object arrays or scalars to perform
hard conversion of objects to a specified type:

• to_numeric() (conversion to numeric dtypes)

  In [370]: m = ['1.1', 2, 3]
  
  In [371]: pd.to_numeric(m)
  Out[371]: array([1.1, 2. , 3. ])
  

• to_datetime() (conversion to datetime objects)

  In [372]: import datetime
  
  In [373]: m = ['2016-07-09', datetime.datetime(2016, 3, 2)]
  
  In [374]: pd.to_datetime(m)
  Out[374]: DatetimeIndex(['2016-07-09', '2016-03-02'], dtype='datetime64[ns]', freq=None)
  

• to_timedelta() (conversion to timedelta objects)

  In [375]: m = ['5us', pd.Timedelta('1day')]
  
  In [376]: pd.to_timedelta(m)
  Out[376]: TimedeltaIndex(['0 days 00:00:00.000005', '1 days 00:00:00'], dtype='timedelta64[ns]', freq=None)
  

To force a conversion, we can pass in an errors argument, which specifies how pandas should deal with elements
that cannot be converted to desired dtype or object. By default, errors='raise', meaning that any errors encountered
will be raised during the conversion process. However, if errors='coerce', these errors will be ignored and pandas
will convert problematic elements to pd.NaT (for datetime and timedelta) or np.nan (for numeric). This might be
useful if you are reading in data which is mostly of the desired dtype (e.g. numeric, datetime), but occasionally has
non-conforming elements intermixed that you want to represent as missing:

In [377]: import datetime

In [378]: m = ['apple', datetime.datetime(2016, 3, 2)]

In [379]: pd.to_datetime(m, errors='coerce')
Out[379]: DatetimeIndex(['NaT', '2016-03-02'], dtype='datetime64[ns]', freq=None)

In [380]: m = ['apple', 2, 3]

In [381]: pd.to_numeric(m, errors='coerce')
Out[381]: array([nan,  2.,  3.])

In [382]: m = ['apple', pd.Timedelta('1day')]

In [383]: pd.to_timedelta(m, errors='coerce')
Out[383]: TimedeltaIndex([NaT, '1 days'], dtype='timedelta64[ns]', freq=None)

The errors parameter has a third option of errors='ignore', which will simply return the passed in data if it
encounters any errors with the conversion to a desired data type:

In [384]: import datetime

In [385]: m = ['apple', datetime.datetime(2016, 3, 2)]

In [386]: pd.to_datetime(m, errors='ignore')
Out[386]: Index(['apple', 2016-03-02 00:00:00], dtype='object')

In [387]: m = ['apple', 2, 3]

In [388]: pd.to_numeric(m, errors='ignore')
Out[388]: array(['apple', 2, 3], dtype=object)

In [389]: m = ['apple', pd.Timedelta('1day')]

In [390]: pd.to_timedelta(m, errors='ignore')
Out[390]: array(['apple', Timedelta('1 days 00:00:00')], dtype=object)

In addition to object conversion, to_numeric() provides another argument downcast, which gives the
option of downcasting the newly (or already) numeric data to a smaller dtype, which can conserve memory:

In [391]: m = ['1', 2, 3]

In [392]: pd.to_numeric(m, downcast='integer')   # smallest signed int dtype
Out[392]: array([1, 2, 3], dtype=int8)

In [393]: pd.to_numeric(m, downcast='signed')    # same as 'integer'
Out[393]: array([1, 2, 3], dtype=int8)

In [394]: pd.to_numeric(m, downcast='unsigned')  # smallest unsigned int dtype
Out[394]: array([1, 2, 3], dtype=uint8)

In [395]: pd.to_numeric(m, downcast='float')     # smallest float dtype
Out[395]: array([1., 2., 3.], dtype=float32)

As these methods apply only to one-dimensional arrays, lists or scalars; they cannot be used directly on multi-dimensional objects such
as DataFrames. However, with apply(), we can “apply” the function over each column efficiently:

In [396]: import datetime

In [397]: df = pd.DataFrame([
   .....:     ['2016-07-09', datetime.datetime(2016, 3, 2)]] * 2, dtype='O')
   .....: 

In [398]: df
Out[398]: 
            0                    1
0  2016-07-09  2016-03-02 00:00:00
1  2016-07-09  2016-03-02 00:00:00

In [399]: df.apply(pd.to_datetime)
Out[399]: 
           0          1
0 2016-07-09 2016-03-02
1 2016-07-09 2016-03-02

In [400]: df = pd.DataFrame([['1.1', 2, 3]] * 2, dtype='O')

In [401]: df
Out[401]: 
     0  1  2
0  1.1  2  3
1  1.1  2  3

In [402]: df.apply(pd.to_numeric)
Out[402]: 
     0  1  2
0  1.1  2  3
1  1.1  2  3

In [403]: df = pd.DataFrame([['5us', pd.Timedelta('1day')]] * 2, dtype='O')

In [404]: df
Out[404]: 
     0                1
0  5us  1 days 00:00:00
1  5us  1 days 00:00:00

In [405]: df.apply(pd.to_timedelta)
Out[405]: 
                0      1
0 00:00:00.000005 1 days
1 00:00:00.000005 1 days

gotchas¶

Performing selection operations on integer type data can easily upcast the data to floating.
The dtype of the input data will be preserved in cases where nans are not introduced.
See also Support for integer NA.

In [406]: dfi = df3.astype('int32')

In [407]: dfi['E'] = 1

In [408]: dfi
Out[408]: 
   A  B    C  E
0  1  0    0  1
1  3  1    0  1
2  0  0  255  1
3  0  1    0  1
4 -1 -1    0  1
5  1  0    0  1
6  0 -1    1  1
7  0  0    0  1

In [409]: dfi.dtypes
Out[409]: 
A    int32
B    int32
C    int32
E    int64
dtype: object

In [410]: casted = dfi[dfi > 0]

In [411]: casted
Out[411]: 
     A    B      C  E
0  1.0  NaN    NaN  1
1  3.0  1.0    NaN  1
2  NaN  NaN  255.0  1
3  NaN  1.0    NaN  1
4  NaN  NaN    NaN  1
5  1.0  NaN    NaN  1
6  NaN  NaN    1.0  1
7  NaN  NaN    NaN  1

In [412]: casted.dtypes
Out[412]: 
A    float64
B    float64
C    float64
E      int64
dtype: object

While float dtypes are unchanged.

In [413]: dfa = df3.copy()

In [414]: dfa['A'] = dfa['A'].astype('float32')

In [415]: dfa.dtypes
Out[415]: 
A    float32
B    float64
C    float64
dtype: object

In [416]: casted = dfa[df2 > 0]

In [417]: casted
Out[417]: 
          A         B      C
0  1.047606  0.256090    NaN
1  3.497968  1.426469    NaN
2       NaN       NaN  255.0
3       NaN  1.139976    NaN
4       NaN       NaN    NaN
5  1.346426  0.096706    NaN
6       NaN       NaN    1.0
7       NaN       NaN    NaN

In [418]: casted.dtypes
Out[418]: 
A    float32
B    float64
C    float64
dtype: object