Different functions over a list of columns and generate new column names automatically with data.table

Question

I have a section in my Shiny app that generates a list.

names of the list are column names of the dataframe we will calculate on, list items contain the ca

Answer 1

If I understand correctly, the question is not about shiny in first place but about how to apply different aggregation functions to specific columns of a data.table.

The names of the columns and the functions which are to be applied on are given as list mylist which is created by the shiny app.

Among the various approaches my preferred option is to compute on the language, i.e., to create a complete expression from the contents of mylist and to evaluate it:

library(magrittr)
library(data.table)
mylist %>%
  names() %>% 
  lapply(
    function(.col) lapply(
      mylist[[.col]], 
      function(.fct) sprintf("%s.%s = %s(%s)", .col, .fct, .fct, .col))) %>% 
  unlist() %>% 
  paste(collapse = ", ") %>% 
  sprintf("as.data.table(mtcars)[, .(%s), by = cyl]", .) %>% 
  parse(text = .) %>% 
  eval()

which yields the expected result

   cyl disp.sum disp.mean    hp.sd drat.sum drat.mean wt.max
1:   6   1283.2  183.3143 24.26049    25.10  3.585714  3.460
2:   4   1156.5  105.1364 20.93453    44.78  4.070909  3.190
3:   8   4943.4  353.1000 50.97689    45.21  3.229286  5.424

The character string which is parsed is created by

mylist %>%
  names() %>% 
  lapply(
    function(.col) lapply(
      mylist[[.col]], 
      function(.fct) sprintf("%s.%s = %s(%s)", .col, .fct, .fct, .col))) %>% 
  unlist() %>% 
  paste(collapse = ", ") %>% 
  sprintf("as.data.table(mtcars)[, .(%s), by = cyl]", .)

and looks as if coded manually:

[1] "as.data.table(mtcars)[, .(disp.sum = sum(disp), disp.mean = mean(disp), hp.sd = sd(hp), drat.sum = sum(drat), drat.mean = mean(drat), wt.max = max(wt)), by = cyl]"

Data

For demonstration, mylist is provided "hard-coded":

mylist <- list(
  disp = c("sum", "mean"),
  hp = "sd",
  drat = c("sum", "mean"),
  wt = "max")