CODEWITHSUNDEEP
r
user11679816
user11679816

Reputation: 31

convert complex for loop to lapply/sapply

Need to convert a for loop to lapply to reduce computational work, however failed to adjust all i labels in the lapply since it is a long process

sample: 


  for (i in 1:76){
    tryCatch({
      model <- nplr(x = splitlist[[i]]$dose, y = splitlist[[i]]$IC, npars = "all")

 .... computation of some variables, then 

      vdata = data.frame(splitlist[[i]]$dose, splitlist[[i]]$viability, splitlist[[i]]$IC)

further computation of variables assingned to i

then:

path = paste0(dir_path,"/results/Patient/",input$txt,"/combined_plots/",titleframe[i, ],".png", sep ="")

  }

...etc

Upvotes: 1

Views: 289

Answers (1)

MDEWITT
MDEWITT

Reputation: 2368

You could try to write a function that does your operation, then lapply said function.

EX

complicated_function <- function(df){
tryCatch({
      model <- nplr(x = df$dose, y = df$IC, npars = "all")
      ---stuff ----
      out <- stuff
}

Then you could do:

outobject <- lapply(your_data, FUN = complicated_function)

I don't know if that puts on you on the right track. If you have some representative data I could help more, but that methodology probably will work for you.

Here is an even more complete example

# Generate Some Fake Data
library(tidyverse)

create_random_dfs <- function(x){
  cats <- sample(LETTERS, size = 10, replace = TRUE)
  x1 <- rnorm(10, 0, 1)
  x2 <- rnorm(10, 0, 1)
  y <- 3*x1 + 1*x2
  tibble(cats, x1, x2, y)
}

df <- lapply(1:100, create_random_dfs)

# Custom Function that works for one dataframe
df[[1]]

my_function <- function(input){
  fit <- lm(y ~ x1 + x2, data = input)
  broom::tidy(fit)
}

lapply(df, my_function)

Upvotes: 1

Related Questions