I am a R novice currently experiencing some difficulty with my code. Essentially, I have several variables in a dataset that contain information on which types of activities an individual frequently participates in (e.g. 1 = reading, 2 = arts and crafts, 3 = gardening, ect..).
Some simulated data:
df = data.frame(ID = c(1001, 1002, 1003, 1004, 1005,1006,1007,1008,1009,1010,1011),
orig_1 = c('-7', '2','1','1','NA','2', '3','NA','NA','2', '2'),
orig_2 = c('1','1','2','1','3','2', '2', '3','NA','2', '2'),
orig_3 = c('-7','3','NA','1','NA','2','NA','1','NA','2', '2'))
Based on these variables, I would like to create new variables that, for instance, reflects whether a person participates in a specific (e.g. 0 = no, 1 = yes). The first thing I have done is code values that correspond to "do not know" as NA:
#Recode variables
df$orig_1[df$orig_1==-7] <- NA
df$orig_2[df$orig_2==-7] <- NA
df$orig_3[df$orig_3==-7] <- NA
Then I have created my new 'activity' variables:
# create new activity variable
df$activity_1 <- NA
df$activity_2 <- NA
df$activity_3 <- NA
Next, I have adapted a function (kindly suggested by @Sonny) to search through the following columns and return a "1" (for those reporting to have participated in an activity) or otherwise a "0":
df$activity_1 <- na.omit(apply(df[, 2:4], 1, function(x) {
if(any(x %in% c(1))) {
return(1)
} else {
return(0)
}
}))
df$activity_2 <- na.omit(apply(df[, 2:4], 1, function(x) {
if(any(x %in% c(2))) {
return(1)
} else {
return(0)
}
}))
df$activity_3 <- na.omit(apply(df[, 2:4], 1, function(x) {
if(any(x %in% c(3))) {
return(1)
} else {
return(0)
}
}))
This part does not work but the idea here was to introduce Na's to the new variables if all original variables were equivalent to "NA":
df$activity_1[df$orig_1==NA & df$orig_2==NA & df$orig_3==NA] <- NA
Ideally, the resulting dataframe should look like this:
ID orig_1 orig_2 orig_3 activity_1 activity_2 activity_3
1 1001 NA 1 NA 1 0 0
2 1002 2 1 NA 1 1 0
3 1003 1 2 NA 1 1 0
4 1004 1 1 1 1 0 0
5 1005 NA 3 NA 0 0 1
6 1006 2 2 2 0 1 0
7 1007 3 2 NA 0 1 1
8 1008 NA NA 1 1 0 0
9 1009 NA NA NA NA NA NA
10 1010 2 2 2 0 1 0
11 1011 2 2 2 0 1 0
I would greatly appreciate any suggestions you may have on improving this code!
First, you need to make true NAs. You're doing 'NA' which is a string and different from NA. We can fix this like so:
df[df == "NA"] <- NA
Then we could look in an apply where all columns 2:4 are NA and set the activity_* columns accordingly.
df[apply(df[2:4], 1, function(x) all(is.na(x))), 5:7] <- NA
Or vectorized, as @akrun suggested:
df[!rowSums(!is.na(df[2:4])), 5:7] <- NA
df
# ID orig_1 orig_2 orig_3 activity_1 activity_2 activity_3
# 1 1001 -7 1 -7 1 0 0
# 2 1002 2 1 3 1 1 1
# 3 1003 1 2 <NA> 1 1 0
# 4 1004 1 1 1 1 0 0
# 5 1005 <NA> 3 <NA> 0 0 1
# 6 1006 2 2 2 0 1 0
# 7 1007 3 2 <NA> 0 1 1
# 8 1008 <NA> 3 1 1 0 1
# 9 1009 <NA> <NA> <NA> NA NA NA
# 10 1010 2 2 2 0 1 0
# 11 1011 2 2 2 0 1 0
Data
df <- structure(list(ID = c(1001, 1002, 1003, 1004, 1005, 1006, 1007,
1008, 1009, 1010, 1011), orig_1 = structure(c(NA, 3L, 2L, 2L,
5L, 3L, 4L, 5L, 5L, 3L, 3L), .Label = c("-7", "1", "2", "3",
"NA"), class = "factor"), orig_2 = structure(c(1L, 1L, 2L, 1L,
3L, 2L, 2L, 3L, 4L, 2L, 2L), .Label = c("1", "2", "3", "NA"), class = "factor"),
orig_3 = structure(c(NA, 4L, 5L, 2L, 5L, 3L, 5L, 2L, 5L,
3L, 3L), .Label = c("-7", "1", "2", "3", "NA"), class = "factor"),
activity_1 = c(1, 1, 1, 1, 0, 0, 0, 1, 0, 0, 0), activity_2 = c(0,
1, 1, 0, 0, 1, 1, 0, 0, 1, 1), activity_3 = c(0, 1, 0, 0,
1, 0, 1, 1, 0, 0, 0)), .Names = c("ID", "orig_1", "orig_2",
"orig_3", "activity_1", "activity_2", "activity_3"), row.names = c(NA,
-11L), class = "data.frame")