I want to replace a value fu_cm[i] to be missing if a value fu_cm_taken[i] and fu_cm_yn[i] are not 1. i ranges from 1-16. I'm trying to use ifelse but getting the following.
unexpected token 'in'
unexpected token 'fu_cm_taken
length_times_renamed <- length_times_renamed %>%
foreach(i in 1: 16) {
if fu_cm_taken[i]!=1 & fu_cm_yn[i]!=1
{(fu_cm[i] = .)
else(fu_cm[i])}}
Are fu_cum, fu_cm_taken, and fu_cm_yn all columns in a data frame named length_times_renamed? If not, what is length_times_renamed? Assuming they are all column names, I would use functions from the dplyr package to do this.
library(dplyr)
length_times_renamed <- length_times_renamed |>
mutate(fu_cm = ifelse(fu_cm_taken == 1 & fu_cm_yn ==1, fu_cm, NA))
Note that the logic is fu_cm_taken and fu_cm_yn both have to equal 1 or fu_cm is set to NA. Your code seemed to set fu_cm to an alternative value if both fu_cm_yn nad fu_cm_taken were not 1.
Among the problems with your code is that you try to pipe length_times_renamed into a loop. That is not a correct use of the pipe operator. Also, foreach is not a function I'm familiar with.
Yes, but how would I incorporate all 16 of these?
Here's a look at what I'm doing now, i[2:5]
> lengths_times_renamed$fu_cm2 <- ifelse(lengths_times_renamed$fu_cm_taken2!=1 & lengths_times_renamed$fu_cm_yn2!=1,NA,lengths_times_renamed$fu_cm2)
> lengths_times_renamed$fu_cm3 <- ifelse(lengths_times_renamed$fu_cm_taken3!=1 & lengths_times_renamed$fu_cm_yn3!=1,NA,lengths_times_renamed$fu_cm3)
> lengths_times_renamed$fu_cm4 <- ifelse(lengths_times_renamed$fu_cm_taken4!=1 & lengths_times_renamed$fu_cm_yn4!=1,NA,lengths_times_renamed$fu_cm4)
> lengths_times_renamed$fu_cm5 <- ifelse(lengths_times_renamed$fu_cm_taken5!=1 & lengths_times_renamed$fu_cm_yn5!=1,NA,lengths_times_renamed$fu_cm5)
Like many R function, ifelse() and mutate() are vectorized, they act on the whole column or vector without the need for a loop. Here is a toy example of using my code on an invented data frame.
length_times_renamed <- data.frame(fu_cm_taken = c(1,1,0,1),
fu_cm_yn = c(1,0,1,1),
fu_cm = c(4,5,6,7))
length_times_renamed
#> fu_cm_taken fu_cm_yn fu_cm
#> 1 1 1 4
#> 2 1 0 5
#> 3 0 1 6
#> 4 1 1 7
library(dplyr)
length_times_renamed <- length_times_renamed |>
mutate(fu_cm = ifelse(fu_cm_taken == 1 & fu_cm_yn ==1, fu_cm, NA))
length_times_renamed
#> fu_cm_taken fu_cm_yn fu_cm
#> 1 1 1 4
#> 2 1 0 NA
#> 3 0 1 NA
#> 4 1 1 7
Created on 2023-06-09 with reprex v2.0.2
My column names are fu_cm_taken2, fu_cm_yn2, fu_cm2, fu_cm_taken3, fu_cm_yn3, fu_cm_taken3... etc.
Please upload the output of
dput(head(length_times_renamed, 10))
Okay, so basically you have a more advanced options imo. This would be to reshape your data and simplify it, to reduce the necessary code. This would involve a few steps (you could reshape you data to long, split the column names to obtain "name + number", then apply a function which applies your desired logic to each subset based on the subset).
Here is the code with a little bit of toy data and I hope I understood the data you have well enough, since you did not provide a full reprex:
set.seed(0)
Data <- data.frame(
fu_cm_taken1 = sample(c(0,1), 50, replace = TRUE),
fu_cm_yn1 = sample(c(0,1), 50, replace= TRUE),
fu_cm1 = sample(c("yay", "ney"), 50, replace = TRUE),
fu_cm_taken2 = sample(c(0,1), 50, replace = TRUE),
fu_cm_yn2 = sample(c(0,1), 50, replace= TRUE),
fu_cm2 = sample(c("yay", "ney"), 50, replace = TRUE)
)
library(data.table)
Data <- as.data.table(Data)
# Reshape to long
Data <- melt.data.table(Data, measure.vars = colnames(Data))
# Split with Regex
Data[, c("Name", "Number") := tstrsplit(variable, "(?<=[a-zA-Z])(?=[[:digit:]])", perl = TRUE)]
# Define function
myFunc <- function(dt) {
dt[, id := 1:.N, Name]
dt |>
dcast.data.table(id ~ Name, value.var = "value") |>
collapse::ftransform(
fu_cm = fifelse(fu_cm_taken != 1 & fu_cm_yn != 1, fu_cm, NA_character_)
) |>
melt.data.table(id.vars = "id") |>
collapse::fselect(- id)
}
# Apply to each subset defined by Number
Data <- Data |>
collapse::rsplit(~ Number) |>
collapse::rapply2d(
myFunc
) |>
collapse::unlist2d() |>
# this returns a data.frame, so convert back, but efficently
collapse::qDT()
# Reshape to original form
Data[, orig_name := paste0(variable, .id)]
Data[, id := 1:.N, .(variable, .id)]
Data[, .id := NULL]
dcast.data.table(Data, id ~ orig_name, value.var = "value")
id fu_cm1 fu_cm2 fu_cm_taken1 fu_cm_taken2 fu_cm_yn1 fu_cm_yn2
1: 1 <NA> ney 1 0 0 0
2: 2 <NA> <NA> 0 1 1 1
3: 3 <NA> <NA> 1 0 0 1
4: 4 <NA> yay 0 0 1 0
5: 5 <NA> <NA> 0 0 1 1
6: 6 <NA> <NA> 1 0 0 1
7: 7 yay <NA> 0 1 0 1
8: 8 <NA> <NA> 0 1 1 0
9: 9 <NA> <NA> 0 1 1 0
10: 10 <NA> <NA> 1 1 1 1
11: 11 <NA> <NA> 1 1 0 0
12: 12 yay ney 0 0 0 0
13: 13 <NA> yay 0 0 1 0
14: 14 <NA> yay 0 0 1 0
15: 15 <NA> <NA> 0 1 1 1
16: 16 <NA> yay 0 0 1 0
17: 17 <NA> ney 1 0 1 0
18: 18 <NA> <NA> 1 1 1 0
19: 19 <NA> <NA> 1 0 0 1
20: 20 <NA> <NA> 1 0 1 1
21: 21 <NA> <NA> 0 1 1 0
22: 22 <NA> <NA> 0 0 1 1
23: 23 <NA> <NA> 0 0 1 1
24: 24 yay <NA> 0 1 0 1
25: 25 yay <NA> 0 0 0 1
26: 26 ney <NA> 0 0 0 1
27: 27 <NA> <NA> 0 1 1 0
28: 28 <NA> <NA> 1 1 1 0
29: 29 ney <NA> 0 1 0 0
30: 30 ney <NA> 0 1 0 0
31: 31 <NA> <NA> 1 0 1 1
32: 32 <NA> <NA> 1 1 1 0
33: 33 <NA> ney 1 0 1 0
34: 34 ney yay 0 0 0 0
35: 35 <NA> ney 1 0 0 0
36: 36 ney ney 0 0 0 0
37: 37 <NA> <NA> 0 0 1 1
38: 38 <NA> <NA> 1 1 0 1
39: 39 <NA> <NA> 0 0 1 1
40: 40 <NA> <NA> 1 1 0 1
41: 41 <NA> <NA> 1 0 1 1
42: 42 <NA> <NA> 1 0 0 1
43: 43 <NA> <NA> 1 1 0 1
44: 44 <NA> ney 0 0 1 0
45: 45 <NA> <NA> 1 1 1 1
46: 46 <NA> <NA> 1 1 0 0
47: 47 <NA> <NA> 1 1 0 1
48: 48 <NA> ney 1 0 0 0
49: 49 <NA> <NA> 1 1 1 1
50: 50 <NA> ney 0 0 1 0
I hope the comments explain the code well enough, if you have questions or don't think this fits your needs, just raise a question and take me. I will try to answer them.
According to the foreach command and the missing value ., I assumed that you might mix some Stata syntax here.
Based on your original code, in R it may like this:
length_times_renamed <- length_times_renamed %>%
for (i in 1:16) {
if (fu_cm_taken[i]!=1 & fu_cm_yn[i]!=1) {
fu_cm[i] = NA
}
else {
fu_cm[i] = fu_cm[i]
}
}
In this case, using ifelse is a better option.
This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.
If you have a query related to it or one of the replies, start a new topic and refer back with a link.