CODEWITHSUNDEEP
rloopsdataframe
Felix
Felix

Reputation: 309

Deleting certain rows without a loop

I have a dataframe with car repair data. It also tells me if a car broke down (and got repaired). I would like to get rid of all rows with repair events after the car broke down.

Car <- c('A','A','B')
Damage <-c("Wheels","Motor","Motor")
date <-c('01-01-2015','01-01-2016','01-01-2016')
Broke_down <-c(1,0,1)

df <- as.data.frame(cbind(Car,date,Damage,Broke_down))

Basically, I want to remove all lines that occur for a car after the dummy was 1.

So in this case the output would have to be:

"Car"        "date"          "Damage"     "Broke_down"
 A           01-01-2015       Wheels         1
 B           01-01-2016       Motor          1

Best, Felix

Upvotes: 1

Views: 74

Answers (3)

snaut
snaut

Reputation: 2535

Here's a solution using split and lapply, data preparation is the same as in the question:

df2 <- do.call(
  rbind,
  lapply(
    split(df, df$Car),
    function(x){
      x[1:which.min(x$Broke_down==1), ]
    })
)

Explanation:
split gives a list of data.frames
lapply applies the function in it's second argument and returns a list of the results
finally do.call calls rbind with the resulting list of data.frames giving you one long data.frame again.

There are similar but faster solutions using data.table and dplyr.

Upvotes: 2

akrun
akrun

Reputation: 887118

Based on the update in OP's post

library(data.table)
setDT(df)[, .SD[cummin(Broke_down) > 0], Car]
#    Car       date Damage Broke_down
#1:   A 01-01-2015 Wheels          1
#2:   B 01-01-2016  Motor          1

Or with ave from base R

df[with(df, ave(Broke_down, Car, FUN = cummin) > 0),]
#  Car       date Damage Broke_down
#1   A 01-01-2015 Wheels          1
#3   B 01-01-2016  Motor          1

Upvotes: 1

Val
Val

Reputation: 7023

There might be a more elegant way, but lapply and do.call do the trick:

df_out <- do.call(rbind,lapply(unique(df$Car),function(x){

  df_sub <- df[df$Car==x,]
  df_sub[1:which(df_sub$Broke_down == 1)[1],]

}))

> df_out
  Car       date Damage Broke_down
1   A 01-01-2015 Wheels          1
3   B 01-01-2016  Motor          1

Upvotes: 0

Related Questions