How to create another table in R to calculate the difference?

Question

I have a set of data frame as below:

ID	Parameter	value
123-01	a1	x
123-02	a1	x
123-01	b3	x
123-02	b3	x
124-01	a1	x
125-01	a1	x
126-01	a1	x
124-01	b3	x
125-01	b3	x
126-01	b3	x

I would like to find the sampleID that ended with "-02", and calculate the difference of the same sample ID that has the same first three digit by same parameter.

For example, calculate the difference of 123-01 and 123-02 based on parameter a1. Then the difference of 123-01 and 123-02 based on parameter b3, etc....

In the end, I can get a table contains

ID	Parameter	DiffValue
123	a1	y
123	b3	y
127	a1	y
127	b3	y

How can I do it?

I tried to use dplyr (filter) to create a table that only contains the duplicate, and then how do I match the origin table and do the calculation?

Answer 1

try to do it this way

library(tidyverse)
df <- read.table(text = "ID    Parameter value
123-01  a1  10
123-02  a1  10
123-01  b3  10
123-02  b3  10
124-01  a1  10
125-01  a1  10
126-01  a1  10
124-01  b3  10
125-01  b3  10
126-01  b3  10", header = T)

df %>% 
  arrange(Parameter, ID) %>% 
  separate(ID, into = c("id_grp", "n"), sep = "-", remove = F) %>% 
  group_by(Parameter, id_grp) %>% 
  mutate(diff_value = c(NA, diff(value))) %>% 
  select(-c(id_grp, n))
#> Adding missing grouping variables: `id_grp`
#> # A tibble: 10 x 5
#> # Groups:   Parameter, id_grp [8]
#>    id_grp ID     Parameter value diff_value
#>    <chr>  <chr>  <chr>     <int>      <int>
#>  1 123    123-01 a1           10         NA
#>  2 123    123-02 a1           10          0
#>  3 124    124-01 a1           10         NA
#>  4 125    125-01 a1           10         NA
#>  5 126    126-01 a1           10         NA
#>  6 123    123-01 b3           10         NA
#>  7 123    123-02 b3           10          0
#>  8 124    124-01 b3           10         NA
#>  9 125    125-01 b3           10         NA
#> 10 126    126-01 b3           10         NA

^{Created on 2021-01-26 by the reprex package (v0.3.0)}