How to replace the first value of each group

user_9958112 0
Viewed 2742

My df is as below:

set.seed(123)
df <- data.frame(x = sample(letters[1:3],20,replace = TRUE),
                 y = sample(1:10,20,replace = TRUE))
df <- df[order(df$x),]

I want to replace the first value of each group with NA. For example:

x y
a NA
a 8
a 1
a 8
b NA
b 3
b 2
b 10
b 8
.
.

I've no problem with get first values but that's not meaningful.

test <- df %>% 
  group_by(x) %>% 
  do(a = head(.$y,1))

Please help next step.

dplyr r
2 Answers

With dplyr we could replace values for the 1st row.

library(dplyr)

df %>% group_by(x) %>% mutate(y = replace(y, 1, NA))

#    x         y
#   <fct> <int>
# 1 a        NA
# 2 a         8
# 3 a         1
# 4 a         8
# 5 a         3
# 6 a         4
# 7 b        NA
# 8 b         6
# 9 b         3
#10 b         2
#....

Or using base R ave

with(df, ave(y, x, FUN = function(i) replace(i, 1, NA)))
#[1] NA  8  1  8  3  4 NA  6  3  2 10  8 NA 10  7  6 10  7  5  3
user_39629140

With ifelse and the use of a temp variable (pos) that counts the position in the grouping: 

library(dplyr)
df %>% 
  group_by(x) %>% 
  mutate(pos = 1:n(), 
         y=ifelse(pos==1, NA, y)) %>% 
  select(-pos) %>% # remove this to see what's happening in the ifelse()
  head(8)
# # A tibble: 8 x 2
# # Groups:   x [2]
# x         y
# <fct> <int>
# 1 a        NA
# 2 a         8
# 3 a         1
# 4 a         8
# 5 a         3
# 6 a         4
# 7 b        NA
# 8 b         6
user_56582700
Related