replace_na

In [1]:

# https://tidyr.tidyverse.org/reference/replace_na.html
%run nb_helpers.py

from datar.all import *

nb_header(replace_na)

# https://tidyr.tidyverse.org/reference/replace_na.html %run nb_helpers.py from datar.all import * nb_header(replace_na)

Try this notebook on binder.

★ replace_na

¶

Replace NA with a value¶

This function can be also used not as a verb. As a function called as
an argument in a verb, data is passed implicitly. Then one could
pass data_or_replace as the data to replace.

Args:¶

  data: The data piped in
  data_or_replace: When called as argument of a verb, this is the
    data to replace. Otherwise this is the replacement.

  replace: The value to replace with
    Can only be a scalar or dict for data frame.
    So replace NA with a list is not supported yet.

Returns:¶

  Corresponding data with NAs replaced

In [2]:

df = tibble(x = c(1, 2, NA), y = c("a", NA, "b"))
df >> replace_na(dict(x = 0, y = "unknown"))

df = tibble(x = c(1, 2, NA), y = c("a", NA, "b")) df >> replace_na(dict(x = 0, y = "unknown"))

Out[2]:

	x	y
	<float64>	<object>
0	1.0	a
1	2.0	unknown
2	0.0	b

In [3]:

df >> mutate(x = replace_na(f.x, 0))

df >> mutate(x = replace_na(f.x, 0))

Out[3]:

	x	y
	<float64>	<object>
0	1.0	a
1	2.0	NaN
2	0.0	b

In [4]:

df.x >> replace_na(0)

df.x >> replace_na(0)

Out[4]:

0    1.0
1    2.0
2    0.0
Name: x, dtype: float64

In [5]:

df.y >> replace_na("unknown")

df.y >> replace_na("unknown")

Out[5]:

0          a
1    unknown
2          b
Name: y, dtype: object

In [6]:

df_list = tibble(z = [seq(1,5), NULL, seq(10,20)])
df_list >> replace_na({'z': 5}) # replace with a list not supported yet

df_list = tibble(z = [seq(1,5), NULL, seq(10,20)]) df_list >> replace_na({'z': 5}) # replace with a list not supported yet

Out[6]:

	z
	<object>
0	[1, 2, 3, 4, 5]
1	5
2	[10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]