Transpose data frame in r - optimizing the code

Question

I have got below data frame in R and found a way to transpose it but I would like to optimize the code.

So my data frame is as this one:

EVENT NUMBER HOURS_PCT CYCLES_PCT
A23   -17    2         2
A23   -18    3         3
A23   -19    4         4
A23   -20    5         5
A23   -21    6         6
A23   -22    7         7
A23   -23    8         8
A23   -24    9         9
A23   -25    10        10
CD12   -1    11        11
CD12   -2    12        12
CD12   -3    13        13
CD12   -4    14        14
CD12   -5    15        15
CD12   -6    16        16
CD12   -7    17        17

The results are fine and correct and look as below:

EVENT  <- c('A23','A23','A23','A23','A23','A23','A23','A23','A23','CD12','CD12','CD12','CD12','CD12','CD12','CD12')
NUMBER <-c('-17','-18','-19','-20','-21','-22','-23','-24','-25','-1','-2','-3','-4','-5','-6','-7')
HOURS_PCT <- seq(from=2, to=17, by=1)
CYCLES_PCT <- seq(from=2, to=17, by=1)

df <- data.frame(EVENT, NUMBER, HOURS_PCT, CYCLES_PCT)

df_1h <- df %>%
  arrange(EVENT, NUMBER, HOURS_PCT) %>%
  group_by(EVENT) %>% top_n(5,NUMBER) %>% 
  mutate(SEQ = row_number())

df_1c <- df %>%
  arrange(EVENT, NUMBER, CYCLES_PCT) %>%
  group_by(EVENT) %>% top_n(5,NUMBER) %>% 
  mutate(SEQ = row_number())

df_1h$NUMBER<-NULL; df_1h$CYCLES_PCT<-NULL
df_1c$NUMBER<-NULL; df_1c$HOURS_PCT<-NULL

df_1h_t <- spread(df_1h, SEQ, HOURS_PCT, fill = "")
df_1c_t <- spread(df_1c, SEQ, CYCLES_PCT, fill = "")

df_final <- cbind(df_1h_t,df_1c_t)
df_final$EVENT1<-NULL

I find it very manual and wonder if it can be optimized. I tried adding gather and spread into my piping commands but they never worked.

Answer 1

I think what you want can be achieved by first transforming the columns which end with "PCT" in long format, select top 5 NUMBER in each EVENT and column, create a unique identifier row and get the data back in wide format.

library(dplyr)
library(tidyr)

df %>%
  pivot_longer(cols = ends_with('PCT')) %>%
  group_by(EVENT, name) %>%
  top_n(5, NUMBER) %>%
  group_by(EVENT) %>%
  mutate(SEQ = row_number()) %>%
  select(-NUMBER, -name)  %>%
  pivot_wider(names_from = SEQ, values_from = value)

#  EVENT   `1`   `2`   `3`   `4`   `5`   `6`   `7`   `8`   `9`  `10`
#  <fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#1 A23       6     6     7     7     8     8     9     9    10    10
#2 CD12     13    13    14    14    15    15    16    16    17    17

---

pivot_longer and pivot_wider are successors of gather and spread . If you haven't updated tidyr yet, using gather and spread you can do

df %>%
  gather(name, value, ends_with('PCT')) %>%
  group_by(EVENT, name) %>%
  top_n(5, NUMBER) %>%
  group_by(EVENT) %>%
  mutate(SEQ = row_number()) %>%
  select(-NUMBER, -name)  %>%
  spread(SEQ, value)