3.6. Series Slicing

3.6.1. SetUp

>>> import pandas as pd

3.6.2. Numeric Index

>>> s = pd.Series([1.0, 2.0, 3.0, 4.0, 5.0])
>>> s
0    1.0
1    2.0
2    3.0
3    4.0
4    5.0
dtype: float64
>>> s[:2]
0    1.0
1    2.0
dtype: float64
>>> s[2:]
2    3.0
3    4.0
4    5.0
dtype: float64
>>> s[1:-2]
1    2.0
2    3.0
dtype: float64
>>> s[::2]
0    1.0
2    3.0
4    5.0
dtype: float64
>>> s[1::2]
1    2.0
3    4.0
dtype: float64

3.6.3. String Index

  • Using string index upper and lower bound are inclusive!

  • String indexes has also numeric index underneath

>>> s = pd.Series(
...     data = [1.0, 2.0, 3.0, 4.0, 5.0],
...     index = ['a', 'b', 'c', 'd', 'e'])
>>> s
a    1.0
b    2.0
c    3.0
d    4.0
e    5.0
dtype: float64
>>> s['a':'d']
a    1.0
b    2.0
c    3.0
d    4.0
dtype: float64
>>> s['a':'d':2]
a    1.0
c    3.0
dtype: float64
>>> s['a':'d':'b']
Traceback (most recent call last):
TypeError: '>=' not supported between instances of 'str' and 'int'
>>> s['d':'a']
Series([], dtype: float64)
>>> s[:2]
a    1.0
b    2.0
dtype: float64
>>> s[2:]
c    3.0
d    4.0
e    5.0
dtype: float64
>>> s[1:-2]
b    2.0
c    3.0
dtype: float64
>>> s[::2]
a    1.0
c    3.0
e    5.0
dtype: float64
>>> s[1::2]
b    2.0
d    4.0
dtype: float64
>>> s = pd.Series(
...     data = [1.0, 2.0, 3.0, 4.0, 5.0],
...     index = ['aaa', 'bbb', 'ccc', 'ddd', 'eee'])
>>>
>>> s
aaa    1.0
bbb    2.0
ccc    3.0
ddd    4.0
eee    5.0
dtype: float64
>>>
>>> s['a':'b']
aaa    1.0
dtype: float64
>>>
>>> s['a':'c']
aaa    1.0
bbb    2.0
dtype: float64

3.6.4. Date Index

>>> s = pd.Series(
...     data = [1.0, 2.0, 3.0, 4.0, 5.0],
...     index = pd.date_range('1999-12-30', periods=5))
>>>
>>> s
1999-12-30    1.0
1999-12-31    2.0
2000-01-01    3.0
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s['2000-01-02':'2000-01-04']
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s['1999-12-30':'2000-01-04':2]
1999-12-30    1.0
2000-01-01    3.0
2000-01-03    5.0
Freq: 2D, dtype: float64
>>> s['1999-12-30':'2000-01-04':-1]
Series([], Freq: -1D, dtype: float64)
>>> s['2000-01-04':'1999-12-30':-1]
2000-01-03    5.0
2000-01-02    4.0
2000-01-01    3.0
1999-12-31    2.0
1999-12-30    1.0
Freq: -1D, dtype: float64
>>> s[:'1999']
1999-12-30    1.0
1999-12-31    2.0
Freq: D, dtype: float64
>>> s['2000':]
2000-01-01    3.0
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s[:'1999-12']
1999-12-30    1.0
1999-12-31    2.0
Freq: D, dtype: float64
>>> s['2000-01':]
2000-01-01    3.0
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s[:'2000-01-02']
1999-12-30    1.0
1999-12-31    2.0
2000-01-01    3.0
2000-01-02    4.0
Freq: D, dtype: float64
>>> s['2000-01-02':]
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s['1999-12':'1999-12']
1999-12-30    1.0
1999-12-31    2.0
Freq: D, dtype: float64
>>> s['2000-01':'2000-01-05']
2000-01-01    3.0
2000-01-02    4.0
2000-01-03    5.0
Freq: D, dtype: float64
>>> s[:'2000-01-05':2]
1999-12-30    1.0
2000-01-01    3.0
2000-01-03    5.0
Freq: 2D, dtype: float64
>>> s[:'2000-01-03':-1]
2000-01-03    5.0
Freq: -1D, dtype: float64

Despite DatetimeIndex, this series also has RangeIndex underneath, which you can slice.

>>> s[1:3]
1999-12-31    2.0
2000-01-01    3.0
Freq: D, dtype: float64
>>>
>>> s[:3]
1999-12-30    1.0
1999-12-31    2.0
2000-01-01    3.0
Freq: D, dtype: float64
>>>
>>> s[:3:2]
1999-12-30    1.0
2000-01-01    3.0
Freq: 2D, dtype: float64
>>>
>>> s[::-1]
2000-01-03    5.0
2000-01-02    4.0
2000-01-01    3.0
1999-12-31    2.0
1999-12-30    1.0
Freq: -1D, dtype: float64

3.6.5. Assignments

Code 3.65. Solution
"""
* Assignment: Series Slice Datetime
* Complexity: easy
* Lines of code: 5 lines
* Time: 3 min

English:
    1. Set random seed to zero
    2. Create `s: pd.Series` with 100 random numbers from standard distribution
    3. Series Index are following dates since 2000
    4. Define `result: pd.Series` with values for dates between 2000-02-14 and end of February 2000
    5. Run doctests - all must succeed

Polish:
    1. Ustaw ziarno losowości na zero
    2. Stwórz `s: pd.Series` z 100 losowymi liczbami z rozkładu normalnego
    3. Indeksem w serii mają być kolejne dni od 2000 roku
    4. Zdefiniuj `result: pd.Series` z wartościami pomiędzy datami od 2000-02-14 do końca lutego 2000
    5. Uruchom doctesty - wszystkie muszą się powieść

Hints:
    * `np.random.randn()`

Tests:
    >>> import sys; sys.tracebacklimit = 0

    >>> pd.set_option('display.width', 500)
    >>> pd.set_option('display.max_columns', 10)
    >>> pd.set_option('display.max_rows', 10)

    >>> assert result is not Ellipsis, \
    'Assign result to variable: `result`'
    >>> assert type(result) is pd.Series, \
    'Variable `result` has invalid type, should be `pd.Series`'

    >>> result  # doctest: +NORMALIZE_WHITESPACE
    2000-02-14   -0.509652
    2000-02-15   -0.438074
    2000-02-16   -1.252795
    2000-02-17    0.777490
    2000-02-18   -1.613898
                    ...
    2000-02-25    0.428332
    2000-02-26    0.066517
    2000-02-27    0.302472
    2000-02-28   -0.634322
    2000-02-29   -0.362741
    Freq: D, Length: 16, dtype: float64
"""

import pandas as pd
import numpy as np
np.random.seed(0)

NUMBER = 100

# type: pd.Series
result = ...

Code 3.66. Solution
"""
* Assignment: Slicing Slice Str
* Complexity: easy
* Lines of code: 10 lines
* Time: 13 min

English:
    1. Create `pd.Series` with 26 random integers in range `[10, 100)`
    2. Name indexes like letters from ASCII alphabet (`ASCII_LOWERCASE: str`)
    3. Find middle letter of alphabet
    4. Slice from series 3 elements up and down from middle
    5. Run doctests - all must succeed

Polish:
    1. Stwórz `pd.Series` z 26 losowymi liczbami całkowitymi z przedziału `<10; 100)`
    2. Nazwij indeksy jak kolejne litery alfabetu ASCII (`ASCII_LOWERCASE: str`)
    3. Znajdź środkową literę alfabetu
    4. Wytnij z serii po 3 elementy w górę i w dół od wyszukanego środka
    5. Uruchom doctesty - wszystkie muszą się powieść

Hints:
    * `np.random.randint(..., ..., size=...)`

Tests:
    >>> import sys; sys.tracebacklimit = 0

    >>> assert result is not Ellipsis, \
    'Assign result to variable: `result`'
    >>> assert type(result) is pd.Series, \
    'Variable `result` has invalid type, should be `pd.Series`'

    >>> result
    j    97
    k    80
    l    98
    m    98
    n    22
    o    68
    p    75
    dtype: int64
"""

from statistics import median_low
import pandas as pd
import numpy as np
np.random.seed(0)


ASCII_LOWERCASE = 'abcdefghijklmnopqrstuvwxyz'

# type: pd.Series
result = ...