Suppress replacement errors in R - r

Question

I have a sequence of data frame subsetting operations. Some of them might fail because the rows to be replaced do not exist. I would still like the others to execute. Example:
source_data[source_data$abbr_d == "bdp",]$party_id <- 32
source_data[source_data$abbr_d == "svp",]$party_id <- 4
source_data[source_data$abbr_d == "cvp",]$party_id <- 2
source_data[source_data$abbr_d == "fdp",]$party_id <- 1
source_data[source_data$abbr_d == "gps",]$party_id <- 13
source_data[source_data$abbr_d == "sp",]$party_id <- 3
source_data[source_data$abbr_d == "csp",]$party_id <- 8
source_data[source_data$abbr_d == "pcs",]$party_id <- 8
Error in `$<-.data.frame`(`*tmp*`, "party_id", value = 13) :
replacement has 1 row, data has 0
source_data[source_data$abbr_d == "lega",]$party_id <- 18
source_data[source_data$abbr_d == "edu",]$party_id <- 16
source_data[source_data$abbr_d == "glp",]$party_id <- 31
I would like the script to continue after the error has been thrown. I've tried using tryCatch() but that doesn't really help because I don't know in advance at which point the replacement will fail.
Is there a way to tell R to just "not care" about those replacement errors? And still continue with the next replacement operations?
The only solution I came up with is to use if-statements like this, which is tedious:
if(nrow(source_data[source_data$abbr_d == "lega", 1]) > 0){
source_data[source_data$abbr_d == "lega",]$party_id <- 18
}
if(nrow(source_data[source_data$abbr_d == "edu", 1]) > 0){
source_data[source_data$abbr_d == "edu",]$party_id <- 16
}
etc...

Answer 1

That is quite verbose code. Luckily, there is a way to get this done in a fraction of the code, and preventing your issue. My suggestion is to use a lookup table to build the party_id column
df = data.frame(abbr_d = sample(LETTERS[1:8], 100, replace = TRUE))
lookup_table = 1:8
names(lookup_table) = LETTERS[1:8]
# A B C D E F G H
# 1 2 3 4 5 6 7 8
df$party_id = lookup_table[df$abbr_d]
So, you create the link between abbr_d and party_id once (here letters and simple numbers, but simply replace your values), and use the df$abbr_d column to subset the lookup table. This maps the labels in abbr_d to the values that correspond to that for party_id.
The error you see is avoided because only addr_d values that are actually in the data are looked up in the lookup table. These unneeded values in the lookup table do not pose an issue.
A dplyr approach as a bonus:
library(dplyr)
df %>% mutate(party_id = lookup_table[abbr_d])

Answer 2

You can use data.table library to mitigate the issue
txt<-"
1,a,1
2,b,2
3,c,3
4,d,4
"
dat = read.delim(textConnection(txt),
header=FALSE,sep=",",strip.white=TRUE)
dat
dat[dat$V2=="e",]$V3<-4
# Error in `$<-.data.frame`(`*tmp*`, "V3", value = 4) :
# le tableau de remplacement a 1 lignes, le tableau remplacé en a 0
library(data.table)
data=as.data.table(dat)
data[data$V2=="e",]$V3<-4
# no error thrown
data.table is often faster than data frame, afaik.