case_when

In [1]:

# https://dplyr.tidyverse.org/reference/case_when.html
%run nb_helpers.py

from datar.data import starwars 
from datar.all import *

nb_header(case_when)

# https://dplyr.tidyverse.org/reference/case_when.html %run nb_helpers.py from datar.data import starwars from datar.all import * nb_header(case_when)

Try this notebook on binder.

★ case_when

¶

Vectorise multiple if_else() statements.¶

Args:¶

  cond: A boolean vector
  value: A vector with values to replace
  *more_cases: A list of tuples (cond, value)

Returns:¶

  A vector with values replaced.

In [2]:

df = tibble(x=range(1,51))
df >> mutate(y=case_when(
    f.x % 35 == 0, "fizz buzz",
    f.x % 5 == 0,  "fizz",
    f.x % 7 == 0,  "buzz",
    True,          as_character(f.x)
)) >> pull(f.y, to="array")

df = tibble(x=range(1,51)) df >> mutate(y=case_when( f.x % 35 == 0, "fizz buzz", f.x % 5 == 0, "fizz", f.x % 7 == 0, "buzz", True, as_character(f.x) )) >> pull(f.y, to="array")

Out[2]:

array(['1', '2', '3', '4', 'fizz', '6', 'buzz', '8', '9', 'fizz', '11',
       '12', '13', 'buzz', 'fizz', '16', '17', '18', '19', 'fizz', 'buzz',
       '22', '23', '24', 'fizz', '26', '27', 'buzz', '29', 'fizz', '31',
       '32', '33', '34', 'fizz buzz', '36', '37', '38', '39', 'fizz',
       '41', 'buzz', '43', '44', 'fizz', '46', '47', '48', 'buzz', 'fizz'],
      dtype=object)

In [3]:

df >> mutate(y=case_when(
    True,          as_character(f.x),
    f.x % 5 == 0,  "fizz",
    f.x % 7 == 0,  "buzz",
    f.x % 35 == 0, "fizz buzz"
)) >> pull(f.y, to="array")

df >> mutate(y=case_when( True, as_character(f.x), f.x % 5 == 0, "fizz", f.x % 7 == 0, "buzz", f.x % 35 == 0, "fizz buzz" )) >> pull(f.y, to="array")

Out[3]:

array(['1', '2', '3', '4', '5', '6', '7', '8', '9', '10', '11', '12',
       '13', '14', '15', '16', '17', '18', '19', '20', '21', '22', '23',
       '24', '25', '26', '27', '28', '29', '30', '31', '32', '33', '34',
       '35', '36', '37', '38', '39', '40', '41', '42', '43', '44', '45',
       '46', '47', '48', '49', '50'], dtype=object)

In [4]:

df >> mutate(y=case_when(
    f.x % 5 == 0,  "fizz",
    f.x % 7 == 0,  "buzz",
    f.x % 35 == 0, "fizz buzz"
)) >> pull(f.y, to="array")

df >> mutate(y=case_when( f.x % 5 == 0, "fizz", f.x % 7 == 0, "buzz", f.x % 35 == 0, "fizz buzz" )) >> pull(f.y, to="array")

Out[4]:

array([nan, nan, nan, nan, 'fizz', nan, 'buzz', nan, nan, 'fizz', nan,
       nan, nan, 'buzz', 'fizz', nan, nan, nan, nan, 'fizz', 'buzz', nan,
       nan, nan, 'fizz', nan, nan, 'buzz', nan, 'fizz', nan, nan, nan,
       nan, 'fizz', nan, nan, nan, nan, 'fizz', nan, 'buzz', nan, nan,
       'fizz', nan, nan, nan, 'buzz', 'fizz'], dtype=object)

In [5]:

df.iloc[[1,2,3], 0] = NA

df >> mutate(y=case_when(
    f.x % 35 == 0, "fizz buzz",
    f.x % 5 == 0,  "fizz",
    f.x % 7 == 0,  "buzz",
    is_na(f.x),    "nope",
    True,          as_character(f.x)
)) >> pull(f.y, to="array")

df.iloc[[1,2,3], 0] = NA df >> mutate(y=case_when( f.x % 35 == 0, "fizz buzz", f.x % 5 == 0, "fizz", f.x % 7 == 0, "buzz", is_na(f.x), "nope", True, as_character(f.x) )) >> pull(f.y, to="array")

Out[5]:

array(['1.0', 'nope', 'nope', 'nope', 'fizz', '6.0', 'buzz', '8.0', '9.0',
       'fizz', '11.0', '12.0', '13.0', 'buzz', 'fizz', '16.0', '17.0',
       '18.0', '19.0', 'fizz', 'buzz', '22.0', '23.0', '24.0', 'fizz',
       '26.0', '27.0', 'buzz', '29.0', 'fizz', '31.0', '32.0', '33.0',
       '34.0', 'fizz buzz', '36.0', '37.0', '38.0', '39.0', 'fizz',
       '41.0', 'buzz', '43.0', '44.0', 'fizz', '46.0', '47.0', '48.0',
       'buzz', 'fizz'], dtype=object)

In [6]:

df >> mutate(y=case_when(
    f.x % 35 == 0, NA,
    f.x % 5 == 0,  "fizz",
    f.x % 7 == 0,  "buzz",
    True,          as_character(f.x)
)) >> pull(f.y, to="array")

df >> mutate(y=case_when( f.x % 35 == 0, NA, f.x % 5 == 0, "fizz", f.x % 7 == 0, "buzz", True, as_character(f.x) )) >> pull(f.y, to="array")

Out[6]:

array(['1.0', 'nan', 'nan', 'nan', 'fizz', '6.0', 'buzz', '8.0', '9.0',
       'fizz', '11.0', '12.0', '13.0', 'buzz', 'fizz', '16.0', '17.0',
       '18.0', '19.0', 'fizz', 'buzz', '22.0', '23.0', '24.0', 'fizz',
       '26.0', '27.0', 'buzz', '29.0', 'fizz', '31.0', '32.0', '33.0',
       '34.0', nan, '36.0', '37.0', '38.0', '39.0', 'fizz', '41.0',
       'buzz', '43.0', '44.0', 'fizz', '46.0', '47.0', '48.0', 'buzz',
       'fizz'], dtype=object)

In [7]:

df >> mutate(y=case_when(
    f.x % 35 == 0, 35,
    f.x % 5 == 0,  5,
    f.x % 7 == 0,  7,
    True,          NA)
) >> pull(f.y, to="array")

df >> mutate(y=case_when( f.x % 35 == 0, 35, f.x % 5 == 0, 5, f.x % 7 == 0, 7, True, NA) ) >> pull(f.y, to="array")

Out[7]:

array([nan, nan, nan, nan,  5., nan,  7., nan, nan,  5., nan, nan, nan,
        7.,  5., nan, nan, nan, nan,  5.,  7., nan, nan, nan,  5., nan,
       nan,  7., nan,  5., nan, nan, nan, nan, 35., nan, nan, nan, nan,
        5., nan,  7., nan, nan,  5., nan, nan, nan,  7.,  5.])

In [8]:

df = tibble(x=seq(-2, 2.1, by=.5))
df >> mutate(y=case_when(
    f.x >= 0, sqrt(f.x),
    True,     f.x
)) >> pull(f.y, to="array")

df = tibble(x=seq(-2, 2.1, by=.5)) df >> mutate(y=case_when( f.x >= 0, sqrt(f.x), True, f.x )) >> pull(f.y, to="array")

RuntimeWarning: invalid value encountered in sqrt

Out[8]:

array([-2.        , -1.5       , -1.        , -0.5       ,  0.        ,
        0.70710678,  1.        ,  1.22474487,  1.41421356])

In [9]:

starwars >> \
  select(f[f.name:f.hair_color], f.gender, f.species) >> \
  mutate(
    type = case_when(
      (f.height > 200) | (f.mass > 200), "large",
      f.species == "Droid"          , "robot",
      True                          , "other"
    )
  )

starwars >> \ select(f[f.name:f.hair_color], f.gender, f.species) >> \ mutate( type = case_when( (f.height > 200) | (f.mass > 200), "large", f.species == "Droid" , "robot", True , "other" ) )

Out[9]:

	name	height	mass	gender	species	type
	<object>	<float64>	<float64>	<object>	<object>	<object>
0	Luke Skywalker	172.0	77.0	masculine	Human	other
1	C-3PO	167.0	75.0	masculine	Droid	robot
2	R2-D2	96.0	32.0	masculine	Droid	robot
3	Darth Vader	202.0	136.0	masculine	Human	large
...	...	...	...	...	...	...
4	Leia Organa	150.0	49.0	feminine	Human	other
82	Rey	NaN	NaN	feminine	Human	other
83	Poe Dameron	NaN	NaN	masculine	Human	other
84	BB8	NaN	NaN	masculine	Droid	robot
85	Captain Phasma	NaN	NaN	NaN	NaN	other
86	Padmé Amidala	165.0	45.0	feminine	Human	other

87 rows × 6 columns

In [10]:

starwars >> \
  mutate(type=case_when(
    (f.height > 200) | (f.mass > 200), "large",
    f.species == "Droid",           "robot",
    True,                           "other"
  )) >> \
  pull(f.type, to="array")

starwars >> \ mutate(type=case_when( (f.height > 200) | (f.mass > 200), "large", f.species == "Droid", "robot", True, "other" )) >> \ pull(f.type, to="array")

Out[10]:

array(['other', 'robot', 'robot', 'large', 'other', 'other', 'other',
       'robot', 'other', 'other', 'other', 'other', 'large', 'other',
       'other', 'large', 'other', 'other', 'other', 'other', 'other',
       'robot', 'other', 'other', 'other', 'other', 'other', 'other',
       'other', 'other', 'other', 'other', 'other', 'other', 'large',
       'large', 'other', 'other', 'other', 'other', 'other', 'other',
       'other', 'other', 'other', 'other', 'other', 'other', 'other',
       'other', 'other', 'other', 'other', 'large', 'other', 'other',
       'other', 'other', 'other', 'other', 'other', 'other', 'other',
       'other', 'other', 'other', 'other', 'other', 'large', 'large',
       'other', 'other', 'robot', 'other', 'other', 'other', 'large',
       'large', 'other', 'other', 'large', 'other', 'other', 'other',
       'robot', 'other', 'other'], dtype=object)

In [ ]: