mutate on selected columns based on name (regular expression)

Here's a sample dataset:

ff <- data.frame(
       id = c(1:4),
       w1...33112..Value = c(10, 20, 30, 40),
       w1...33112..Time = c(4, 3, 2, 1),
       w2...33113..Value = c(1, .9, .75, .7),
       w2...33113..Time = c(10, 50, 30, 20),
       w3...33552..Value = c(1, 2, 3, 4),
       w3...33552..Time = c(.5, .5, .9, .9),
       w4...33442..Value = c(100, 50, 40, 30),
       w4...33442..Time = c(2, 1, 4, 3),
       w5...35692..Value = c(.5, .6, .7, .8)
       )

I want to perform some simple operations (usually using diff) on columns selected based on name - the column name must contain the string Value

The example below is for two variables, and there are dozens of such cases in real data.

ff.2 <- ff %>% mutate (
  w1.used = c(0, diff(w1...33112..Value)), 
  w2.used = c(0, diff(w2...33113..Value)), 
)

The name of the new column should start with a string of characters up to the first dot and the selected string (for example "used").

You could simply do:

ff %>%
  mutate(across(ends_with('Value'), ~c(0, diff(.)),
                .names = "{str_extract(.col, 'w[0-9]+')}.used"))

The mutate_at() function allows you to create new variables based on existing variables that match certain criteria, like if their name contains a certain string. Also, str_extract() function will extract the string until the first dot to create new column names.

Here is the sample code which does exactly what you want:

library(tidyverse)

# Select only columns with "Value" in their names
value_cols <- grep("Value", names(ff), value = TRUE)

# Create new columns
ff <- ff %>%
  mutate_at(
    .vars = value_cols,
    .funs = list(
      used = ~ c(0, diff(.))
    )
  )

# Rename new columns
new_col_names <- str_extract(value_cols, "^[^.]*") %>% paste0(".used")

names(ff)[grep("used", names(ff))] <- new_col_names

ff

A base R option

d <- ff[endsWith(names(ff), "Value")]
u <- d - rbind(d[1, ], d[-nrow(d), ])
ff.2 <- cbind(ff, setNames(u, sub("\\..*", ".used", names(u))))

gives

> ff.2
  id w1...33112..Value w1...33112..Time w2...33113..Value w2...33113..Time
1  1                10                4              1.00               10
2  2                20                3              0.90               50
3  3                30                2              0.75               30
4  4                40                1              0.70               20
  w3...33552..Value w3...33552..Time w4...33442..Value w4...33442..Time
1                 1              0.5               100                2
2                 2              0.5                50                1
3                 3              0.9                40                4
4                 4              0.9                30                3
  w5...35692..Value w1.used w2.used w3.used w4.used w5.used
1               0.5       0    0.00       0       0     0.0
2               0.6      10   -0.10       1     -50     0.1
3               0.7      10   -0.15       1     -10     0.1
4               0.8      10   -0.05       1     -10     0.1