lead-lag

In [1]:

# https://dplyr.tidyverse.org/reference/lead-lag.html
%run nb_helpers.py

from datar.all import *

nb_header(lead, lag, book='lead-lag')

# https://dplyr.tidyverse.org/reference/lead-lag.html %run nb_helpers.py from datar.all import * nb_header(lead, lag, book='lead-lag')

Try this notebook on binder.

★ lead

¶

Shift a vector by n positions.¶

The original API:
https://dplyr.tidyverse.org/reference/lead.html

Args:¶

  x: A vector
  n: The number of positions to shift.
  default: The default value to use for positions that don't exist.
  order_by: A vector of column names to order by.

Returns:¶

  A vector

★ lag

¶

Shift a vector by n positions.¶

The original API:
https://dplyr.tidyverse.org/reference/lag.html

Args:¶

  x: A vector
  n: The number of positions to shift.
  default: The default value to use for positions that don't exist.
  order_by: A vector of column names to order by.

Returns:¶

  A vector

In [2]:

x = [1,2,3,4,5]

lag(x)

x = [1,2,3,4,5] lag(x)

Out[2]:

0    NaN
1    1.0
2    2.0
3    3.0
4    4.0
dtype: float64

In [3]:

lead(x)

lead(x)

Out[3]:

0    2.0
1    3.0
2    4.0
3    5.0
4    NaN
dtype: float64

In [4]:

tibble(behind=lag(x), x=x, ahead=lead(x))

tibble(behind=lag(x), x=x, ahead=lead(x))

Out[4]:

	behind	x	ahead
	<float64>	<int64>	<float64>
0	NaN	1	2.0
1	1.0	2	3.0
2	2.0	3	4.0
3	3.0	4	5.0
4	4.0	5	NaN

In [5]:

lag(x, n=1)

lag(x, n=1)

Out[5]:

0    NaN
1    1.0
2    2.0
3    3.0
4    4.0
dtype: float64

In [6]:

lag(x, n=2)

lag(x, n=2)

Out[6]:

0    NaN
1    NaN
2    1.0
3    2.0
4    3.0
dtype: float64

In [7]:

lead(x, n=1)

lead(x, n=1)

Out[7]:

0    2.0
1    3.0
2    4.0
3    5.0
4    NaN
dtype: float64

In [8]:

lead(x, n=2)

lead(x, n=2)

Out[8]:

0    3.0
1    4.0
2    5.0
3    NaN
4    NaN
dtype: float64

In [9]:

lag(x, default=0)

lag(x, default=0)

Out[9]:

0    0
1    1
2    2
3    3
4    4
dtype: int64

In [10]:

lead(x, default=6)

lead(x, default=6)

Out[10]:

0    2
1    3
2    4
3    5
4    6
dtype: int64

In [11]:

scrambled = slice_sample(
    tibble(year=[2000, 2001, 2002, 2003, 2004, 2005], 
           value=[a**2 for a in range(6)]), 
    prop=1
) 

scrambled >> mutate(previous_year_value = lag(f.value)) >> arrange(f.year)

scrambled = slice_sample( tibble(year=[2000, 2001, 2002, 2003, 2004, 2005], value=[a**2 for a in range(6)]), prop=1 ) scrambled >> mutate(previous_year_value = lag(f.value)) >> arrange(f.year)

Out[11]:

	year	value	previous_year_value
	<int64>	<int64>	<float64>
0	2000	0	1.0
1	2001	1	9.0
2	2002	4	NaN
3	2003	9	4.0
4	2004	16	0.0
5	2005	25	16.0

In [12]:

# Use this for lag(value, order_by = year) instead
scrambled >> mutate(previous_year_value = lag(f.value, order_by=f.year)) >> arrange(f.year)

# Use this for lag(value, order_by = year) instead scrambled >> mutate(previous_year_value = lag(f.value, order_by=f.year)) >> arrange(f.year)

Out[12]:

	year	value	previous_year_value
	<int64>	<int64>	<float64>
0	2000	0	0.0
1	2001	1	NaN
2	2002	4	4.0
3	2003	9	1.0
4	2004	16	9.0
5	2005	25	16.0

In [ ]: