CODEWITHSUNDEEP
r
Rene Chan
Rene Chan

Reputation: 985

How to subset and apply a function across the dataset

I have a list of prices for different items in the same dataset.

abc1 <- c("2005-09-18", "ABC", 99.00)
abc2 <- c("2005-09-19", "ABC", 98.00)
abc3 <- c("2005-09-20", "ABC", 98.50)
abc4 <- c("2005-09-21", "ABC", 97.75)
def1 <- c("2005-09-14", "DEF", 79.00)
def2 <- c("2005-09-15", "DEF", 78.00)
def3 <- c("2005-09-16", "DEF", 78.50)
def4 <- c("2005-09-20", "DEF", 77.75)

df <- data.frame(rbind(abc1, abc2, abc3, abc4, def1, def2, def3, def4))

the above quick table would result in : 

             X1  X2    X3
abc1 2005-09-18 ABC    99
abc2 2005-09-19 ABC    98
abc3 2005-09-20 ABC  98.5
abc4 2005-09-21 ABC 97.75
def1 2005-09-14 DEF    79
def2 2005-09-15 DEF    78
def3 2005-09-16 DEF  78.5
def4 2005-09-20 DEF 77.75

I would like to add a column, say X4, which would be the variation of today, versus the previous day, for a specific X2. So x4 would have the following value:

 X4
 0,0%
-1,0%
 0,5%
-0,8%
 0,0%
-1,3%
 0,6%
-1,0%

The goal would be to do that for all the different items in X3. Ideally without splitting the table. I think the date is always going to be in the right order, but just in case.

Upvotes: 2

Views: 72

Answers (2)

DSGym
DSGym

Reputation: 2867

Just a little hint: You wanted to calculate the difference in percent, not the absolute difference.

You have to adjust the formula to do so, otherwise your results are wrong :-).

df %>%
  dplyr::group_by(X2) %>%
  dplyr::mutate(X4 = (X3/lag(X3, default = first(X3)) - 1) * 100)

  X1         X2       X3     X4
  <fct>      <fct> <dbl>  <dbl>
1 2005-09-18 ABC    99    0    
2 2005-09-19 ABC    98   -1.01 
3 2005-09-20 ABC    98.5  0.510
4 2005-09-21 ABC    97.8 -0.761
5 2005-09-14 DEF    79    0    
6 2005-09-15 DEF    78   -1.27 
7 2005-09-16 DEF    78.5  0.641
8 2005-09-20 DEF    77.8 -0.955

Upvotes: 2

akrun
akrun

Reputation: 887138

We can group by 'X2' and take the difference of adjacent elements with diff

library(dplyr)
df %>%
   group_by(X2) %>%
   mutate(X4 = c(0, diff(X3)))

Or after grouping by 'X2', take the difference between the 'X2' and the lag of 'X2'

df %>%
   group_by(X2) %>%
   mutate(X4 = X3 - lag(X3, default = first(X3)))

Upvotes: 3

Related Questions