assume this is my dataframe ( already created by a sequence of pipes )
df <- data.frame(month = c("May", "May","May","June","June","June", "July", "July","July"),
rate = rep(c(0.3, 0.4, 0.5),times=3),
a = c(5,4,3,10,9,8,20,19,18),
b = c(100,200,300,1,2,3,10,20,30))
df
month rate a b
1 May 0.3 5 100
2 May 0.4 4 200
3 May 0.5 3 300
4 June 0.3 10 1
5 June 0.4 9 2
6 June 0.5 8 3
7 July 0.3 20 10
8 July 0.4 19 20
9 July 0.5 18 30
I am looking to follow the dplyr pipes and create dataframe below:
month r0.3_a r0.3_b r0.4_a r0.4_b r0.5_a r0.5_b
1 May 5 100 4 200 3 300
2 June 10 1 9 2 8 3
3 July 20 10 19 20 18 30
so this is what I did:
library(tidyverse)
library(dplyr)
library(gtools)
df_a <- df %>%
select(-b) %>%
pivot_wider(names_from = rate,
values_from = a,
names_glue = "r{rate}_{.value}" )
df_b <- df %>%
select(-a) %>%
pivot_wider(names_from = rate,
values_from = b,
names_glue = "r{rate}_{.value}" )
output <- inner_join(df_a, df_b, by= "month") %>%
select(month, mixedsort(names(.)[-1]))
My question is: assuming df is coming from a sequence of pipes, is there a way to achieve the output without having to stop the pipes and create two separate df s out of the original df?
>Solution :
Use values_from as a vector of columns
library(tidyr)
library(dplyr)
pivot_wider(df, names_from = rate, values_from = c(a, b),
names_glue = "r{rate}_{.value}") %>%
select(month, gtools::mixedsort(names(.)[-1]))
-output
# A tibble: 3 Ă— 7
month r0.3_a r0.3_b r0.4_a r0.4_b r0.5_a r0.5_b
<chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 May 5 100 4 200 3 300
2 June 10 1 9 2 8 3
3 July 20 10 19 20 18 30