expand

In [1]:

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

from datar.all import *

nb_header(expand, nesting, crossing)

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

Try this notebook on binder.

★ expand

¶

Generates all combination of variables found in a dataset.¶

Args:¶

  data: A data frame
  *args: and,
  **kwargs: columns to expand. Columns can be atomic lists.
    - To find all unique combinations of x, y and z, including       those not present in the data, supply each variable as a
      separate argument: expand(df, x, y, z).

    - To find only the combinations that occur in the data, use       nesting: expand(df, nesting(x, y, z)).

    - You can combine the two forms. For example,       expand(df, nesting(school_id, student_id), date) would
      produce a row for each present school-student combination
      for all possible dates.

  _name_repair: treatment of problematic column names:
    - "minimal": No name repair or checks, beyond basic existence,

    - "unique": Make sure names are unique and not empty,

    - "check_unique": (default value), no name repair,       but check they are unique,

    - "universal": Make the names unique and syntactic

    - a function: apply custom name repair

Returns:¶

  A data frame with all combination of variables.

★ nesting

¶

A helper that only finds combinations already present in the data.¶

Args:¶

  *args: and,
  **kwargs: columns to expand. Columns can be atomic lists.
    - To find all unique combinations of x, y and z, including       those not present in the data, supply each variable as a
      separate argument: expand(df, x, y, z).

    - To find only the combinations that occur in the data, use       nesting: expand(df, nesting(x, y, z)).

    - You can combine the two forms. For example,       expand(df, nesting(school_id, student_id), date) would
      produce a row for each present school-student combination
      for all possible dates.

  _name_repair: treatment of problematic column names:
    - "minimal": No name repair or checks, beyond basic existence,

    - "unique": Make sure names are unique and not empty,

    - "check_unique": (default value), no name repair,       but check they are unique,

    - "universal": Make the names unique and syntactic

    - a function: apply custom name repair

Returns:¶

  A data frame with all combinations in data.

★ crossing

¶

A wrapper around expand_grid() that de-duplicates and sorts its inputs¶

When values are not specified by literal list, they will be sorted.

Args:¶

  *args: and,
  **kwargs: columns to expand. Columns can be atomic lists.
    - To find all unique combinations of x, y and z, including       those not present in the data, supply each variable as a
      separate argument: expand(df, x, y, z).

    - To find only the combinations that occur in the data, use       nesting: expand(df, nesting(x, y, z)).

    - You can combine the two forms. For example,       expand(df, nesting(school_id, student_id), date) would
      produce a row for each present school-student combination
      for all possible dates.

  _name_repair: treatment of problematic column names:
    - "minimal": No name repair or checks, beyond basic existence,

    - "unique": Make sure names are unique and not empty,

    - "check_unique": (default value), no name repair,       but check they are unique,

    - "universal": Make the names unique and syntactic

    - a function: apply custom name repair

Returns:¶

  A data frame with values deduplicated and sorted.

In [2]:

fruits = tibble(
  type   = c("apple", "orange", "apple", "orange", "orange", "orange"),
  year   = c(2010, 2010, 2012, 2010, 2010, 2012),
  size  =  factor(
    c("XS", "S",  "M", "S", "S", "M"),
    levels = c("XS", "S", "M", "L")
  ),
  weights = rnorm(6)
)
fruits

fruits = tibble( type = c("apple", "orange", "apple", "orange", "orange", "orange"), year = c(2010, 2010, 2012, 2010, 2010, 2012), size = factor( c("XS", "S", "M", "S", "S", "M"), levels = c("XS", "S", "M", "L") ), weights = rnorm(6) ) fruits

Out[2]:

	type	year	size	weights
	<object>	<int64>	<category>	<float64>
0	apple	2010	XS	2.522472
1	orange	2010	S	0.206341
2	apple	2012	M	-0.667409
3	orange	2010	S	0.887561
4	orange	2010	S	-1.317738
5	orange	2012	M	-0.228718

In [3]:

fruits >> expand(f.type)

fruits >> expand(f.type)

Out[3]:

	type
	<object>
0	apple
1	orange

In [4]:

fruits >> expand(f.type, f.size)

fruits >> expand(f.type, f.size)

Out[4]:

	type	size
	<object>	<category>
0	apple	XS
1	apple	S
2	apple	M
3	apple	L
4	orange	XS
5	orange	S
6	orange	M
7	orange	L

In [5]:

fruits >> expand(f.type, f.size, f.year)

fruits >> expand(f.type, f.size, f.year)

Out[5]:

	type	size	year
	<object>	<category>	<int64>
0	apple	XS	2010
1	apple	XS	2012
2	apple	S	2010
3	apple	S	2012
4	apple	M	2010
5	apple	M	2012
6	apple	L	2010
7	apple	L	2012
8	orange	XS	2010
9	orange	XS	2012
10	orange	S	2010
11	orange	S	2012
12	orange	M	2010
13	orange	M	2012
14	orange	L	2010
15	orange	L	2012

In [6]:

fruits >> expand(nesting(f.type))

fruits >> expand(nesting(f.type))

Out[6]:

	type
	<object>
0	apple
1	orange

In [7]:

fruits >> expand(nesting(f.type, f.size))

fruits >> expand(nesting(f.type, f.size))

Out[7]:

	type	size
	<object>	<category>
0	apple	XS
1	orange	S
2	apple	M
3	orange	M

In [8]:

fruits >> expand(nesting(f.type, f.size, f.year))

fruits >> expand(nesting(f.type, f.size, f.year))

Out[8]:

	type	size	year
	<object>	<category>	<int64>
0	apple	XS	2010
1	orange	S	2010
2	apple	M	2012
3	orange	M	2012

In [9]:

fruits >> expand(f.type, f.size, full_seq(f.year, 1))

fruits >> expand(f.type, f.size, full_seq(f.year, 1))

Out[9]:

	type	size	_VAR_2
	<object>	<category>	<int64>
0	apple	XS	2010
1	apple	XS	2011
2	apple	XS	2012
3	apple	S	2010
4	apple	S	2011
5	apple	S	2012
6	apple	M	2010
7	apple	M	2011
8	apple	M	2012
9	apple	L	2010
10	apple	L	2011
11	apple	L	2012
12	orange	XS	2010
13	orange	XS	2011
14	orange	XS	2012
15	orange	S	2010
16	orange	S	2011
17	orange	S	2012
18	orange	M	2010
19	orange	M	2011
20	orange	M	2012
21	orange	L	2010
22	orange	L	2011
23	orange	L	2012

In [10]:

fruits >> expand(f.type, f.size, seq(2010, 2012))

fruits >> expand(f.type, f.size, seq(2010, 2012))

Out[10]:

	type	size	_VAR_2
	<object>	<category>	<int64>
0	apple	XS	2010
1	apple	XS	2011
2	apple	XS	2012
3	apple	S	2010
4	apple	S	2011
5	apple	S	2012
6	apple	M	2010
7	apple	M	2011
8	apple	M	2012
9	apple	L	2010
10	apple	L	2011
11	apple	L	2012
12	orange	XS	2010
13	orange	XS	2011
14	orange	XS	2012
15	orange	S	2010
16	orange	S	2011
17	orange	S	2012
18	orange	M	2010
19	orange	M	2011
20	orange	M	2012
21	orange	L	2010
22	orange	L	2011
23	orange	L	2012

In [11]:

fruits >> expand(f.type, f.size, year=seq(2010, 2012))

fruits >> expand(f.type, f.size, year=seq(2010, 2012))

Out[11]:

	type	size	year
	<object>	<category>	<int64>
0	apple	XS	2010
1	apple	XS	2011
2	apple	XS	2012
3	apple	S	2010
4	apple	S	2011
5	apple	S	2012
6	apple	M	2010
7	apple	M	2011
8	apple	M	2012
9	apple	L	2010
10	apple	L	2011
11	apple	L	2012
12	orange	XS	2010
13	orange	XS	2011
14	orange	XS	2012
15	orange	S	2010
16	orange	S	2011
17	orange	S	2012
18	orange	M	2010
19	orange	M	2011
20	orange	M	2012
21	orange	L	2010
22	orange	L	2011
23	orange	L	2012

In [12]:

all = fruits >> expand(f.type, f.size, f.year)
all

all = fruits >> expand(f.type, f.size, f.year) all

Out[12]:

	type	size	year
	<object>	<category>	<int64>
0	apple	XS	2010
1	apple	XS	2012
2	apple	S	2010
3	apple	S	2012
4	apple	M	2010
5	apple	M	2012
6	apple	L	2010
7	apple	L	2012
8	orange	XS	2010
9	orange	XS	2012
10	orange	S	2010
11	orange	S	2012
12	orange	M	2010
13	orange	M	2012
14	orange	L	2010
15	orange	L	2012

In [13]:

all >> anti_join(fruits)

all >> anti_join(fruits)

Out[13]:

	type	size	year
	<object>	<category>	<int64>
1	apple	XS	2012
2	apple	S	2010
3	apple	S	2012
4	apple	M	2010
6	apple	L	2010
7	apple	L	2012
8	orange	XS	2010
9	orange	XS	2012
13	orange	S	2012
14	orange	M	2010
16	orange	L	2010
17	orange	L	2012

In [14]:

fruits >> right_join(all)

fruits >> right_join(all)

Out[14]:

	type	year	size	weights
	<object>	<int64>	<category>	<float64>
0	apple	2010	XS	2.522472
1	apple	2012	XS	NaN
2	apple	2010	S	NaN
3	apple	2012	S	NaN
4	apple	2010	M	NaN
5	apple	2012	M	-0.667409
6	apple	2010	L	NaN
7	apple	2012	L	NaN
8	orange	2010	XS	NaN
9	orange	2012	XS	NaN
10	orange	2010	S	0.206341
11	orange	2010	S	0.887561
12	orange	2010	S	-1.317738
13	orange	2012	S	NaN
14	orange	2010	M	NaN
15	orange	2012	M	-0.228718
16	orange	2010	L	NaN
17	orange	2012	L	NaN