CODEWITHSUNDEEP
rrounding
nba2020
nba2020

Reputation: 688

Conditional rounding with IF statement

I'm trying to round numeric values in a data frame to closest interval. I want to round at different intervals based on how big the number is. I've started with this (coming from excel mindset) but I'm stuck to translate it to R code. Note round_any rounds a number to the closest interval(e.g. 5.13->5, 5.85->6)

library(plyr)    
DataFrame <- sapply(DataFrame, function(x) {
       if(x>1) round_any(x,0.25),
       if(x>5) round_any(x,0.5),
       if(x>10) round_any(x,1),
       else x})

Could you please help me out?

Upvotes: 1

Views: 1369

Answers (2)

nba2020
nba2020

Reputation: 688

Thank you all for your help. Based on your responses the following code worked for my data frame

library(plyr)
library(dplyr)
DataFrame[] <- lapply(DataFrame, function(x){
  round_any(x,  
            case_when(
              x > 10 ~ 1.0,
              x > 5 ~ 0.50,
              x > 1 ~ 0.25,
              TRUE ~ 0.001))})

Upvotes: 0

Mikko Marttila
Mikko Marttila

Reputation: 11878

When using sapply on a data frame, you are iterating over the column vectors rather than individual values. As such, you should be looking at vectorized conditional logic functions: just using the standard if control flow isn't terribly useful, as it can only take scalar (length 1) conditions.

In this case, plyr::round_any can take a vector as the accuracy argument; the dplyr function case_when could be useful here. From ?case_when:

This function allows you to vectorise multiple if and else if statements. It is an R equivalent of the SQL CASE WHEN statement.

Here's an example for the case of a single vector to be rounded:

set.seed(11)

# Generate some raw numbers
x <- runif(8, max = 20)
print(x, digits = 4)
#> [1]  5.54500  0.01037 10.21217  0.28096  1.29380 19.09698  1.72992  5.79950

# Round to differing accuracy
plyr::round_any(
  x,
  dplyr::case_when(
    x > 10 ~ 1.0,
    x > 5 ~ 0.50,
    x > 1 ~ 0.25,
    TRUE ~ 0.001
  )
)
#> [1]  5.500  0.010 10.000  0.281  1.250 19.000  1.750  6.000

Created on 2018-05-11 by the reprex package (v0.2.0).

Upvotes: 2

Related Questions