To create graphs from my data that is currently in the wide format, I am trying to pivot this data to a longer format. Currently the data is like this:
| type | parameter_d1 | parameter_d2 |
|---|---|---|
| 0 | 4 | 8 |
| 1 | 6 | 5 |
In reality there is many different parameters, but they all have the suffix _d1 and _d2.
Finally I would like to have a dataframe like the following:
| type | day | parameter |
|---|---|---|
| 0 | 1 | 4 |
| 0 | 2 | 8 |
| 1 | 1 | 6 |
| 1 | 2 | 5 |
I think this shouldn't be too hard to achieve but I can't find an example doing this anywhere.
CodePudding user response：
You can use matching groups and regular expression in the names_pattern argument in pivot_longer to do so. See the example below that I adjusted from a previous answer, where the parameters are y or se, and days are numbers:
library("tidyverse")
# Some data
raw_data <- data.frame(x = seq(10),
y_d1 = sample(x=20,size=10), # c(1.1, 2.4, 3.5, 4.1, 5.9, 6.7, 7.1, 8.3, 9.4, 10.0)
y_d2 = sample(x=20,size=10),
y_d3 = sample(x=20,size=10),
y_d4 = sample(x=20,size=10),
se_d1 = runif(n=10,min=0,max=1),
se_d2 = runif(n=10,min=0,max=1),
se_d3 = runif(n=10,min=0,max=1),
se_d4 = runif(n=10,min=0,max=1))
# Convert to a long format
raw_data %>%
pivot_longer(!x,
names_pattern = "([[:alpha:]] )_d([0-9])",
names_to = c("parameter", "day"))
#> # A tibble: 80 x 4
#> x parameter day value
#> <int> <chr> <chr> <dbl>
#> 1 1 y 1 12
#> 2 1 y 2 1
#> 3 1 y 3 12
#> 4 1 y 4 9
#> 5 1 se 1 0.840
#> 6 1 se 2 0.650
#> 7 1 se 3 0.523
#> 8 1 se 4 0.211
#> 9 2 y 1 15
#> 10 2 y 2 6
#> # ... with 70 more rows
Created on 2022-10-03 by the reprex package (v2.0.1)
CodePudding user response：
Here's a way with pivot_longer:
library(tidyr)
df %>%
pivot_longer(-type, names_sep = "_", names_to = c(".value", "day"),
names_transform = list(day = readr::parse_number))
type day parameter
1 0 1 4
2 0 2 8
3 1 1 6
4 1 2 5