Make unique strings in a data frame - r

Question

I'd like to make a column of (possibly) non-unique strings into a column of unique strings.
For instance, consider:
df <- data.frame(
'Initials' = c("AA","AB","AB")
, 'Data' = c(1,2,3)
)
df
Initials Data
1 AA 1
2 AB 2
3 AB 3
I would like to obtain this:
Initials Data
1 AA 1
2 AB (1) 2
3 AB (2) 3
Note: I know I could use the rownames to uniquely identify the row, but I'd like to retain the string stored in the Initials column, with a number appended.

Answer 1

transform(df, Initials = ave(as.character(Initials), Initials,
FUN = function(x) if (length(x) > 1) paste0(x, " (", seq(x), ")") else x))
# Initials Data
# 1 AA 1
# 2 AB (1) 2
# 3 AB (2) 3

Answer 2

w <- ave(df$Data, df$Initials, FUN = seq_along )
> df$Initials <- paste(df$Initials, "(", w, ")", sep = "")
# > df
# Initials Data
# 1 AA(1) 1
# 2 AB(1) 2
# 3 AB(2) 3

Answer 3

TLDR
Can use make_unique from the makeunique package (Disclaimer: I'm the author)
# install.packages('makeunique')
library(makeunique)
df <- data.frame(
Initials = c("AA","AB","AB"),
Data = c(1,2,3)
)
df[['Initials']] <- make_unique(df[['Initials']])
df
#> Initials Data
#> 1 AA 1
#> 2 AB (1) 2
#> 3 AB (2) 3
Created on 2022-10-15 by the reprex package (v2.0.1)
Full Answer
Disclaimer: I'm the author the makeunique package
The problem with #Sven Hohenstein's answer is that it can still produce duplicates. For example if your input dataset happens to include a string 'AB (2)'. See example of issue below
df <- data.frame(
Initials = c("AA","AB","AB", "AB (2)"),
Data = c(1,2,3,4)
)
transform(df, Initials = ave(as.character(Initials), Initials,
FUN = function(x) if (length(x) > 1) paste0(x, " (", seq(x), ")") else x))
#> Initials Data
#> 1 AA 1
#> 2 AB (1) 2
#> 3 AB (2) 3
#> 4 AB (2) 4
Created on 2022-10-15 by the reprex package (v2.0.1)
Really you'd either want something like make_unique from the makeunique package that warns you when this happens and lets you change suffix formatting to fix the issue
# install.packages('makeunique')
library(makeunique)
df <- data.frame(
Initials = c("AA","AB","AB", "AB (2)"),
Data = c(1,2,3,4)
)
df[['Initials']] <- make_unique(df[['Initials']])
#> Error in make_unique(df[["Initials"]]): make_unique failed to make vector unique.
#> This is because appending ' <dup_number>' to duplicate values led tocreation of term(s) that were in the original dataset:
#> [AB (2)]
#>
#> Please try again with a different argument for either `wrap_in_brackets` or `sep`
# Change suffix format to fix problem
df[['Initials']] <- make_unique(df[['Initials']], sep = "-")
Created on 2022-10-15 by the reprex package (v2.0.1)
Or, you could make.unique from base R that ensures uniqueness but doesn't let you control how your suffixes look.
If you don't want to use a package, feel free to just copy the source function for make_unique into your code and use it like your own function
make_unique <- function(x, sep = " ", wrap_in_brackets = TRUE, warn_about_type_conversion = TRUE){
if(!(is.character(sep) & length(sep) == 1)) stop('`sep` must be a string, not a ', paste0(class(sep), collapse = " "))
if(!(is.logical(wrap_in_brackets) & length(wrap_in_brackets) == 1)) stop('`wrap_in_brackets` must be a flag, not a ', paste0(class(wrap_in_brackets), collapse = " "))
if(!(is.logical(warn_about_type_conversion) & length(warn_about_type_conversion) == 1)) stop('`warn_about_type_conversion` must be a flag, not a ', paste0(class(warn_about_type_conversion), collapse = " "))
if(!any(is.numeric(x),is.character(x),is.factor(x))) stop('input to `make_unique` must be a character, numeric, or factor variable')
if(is.factor(x)) {
if(warn_about_type_conversion) warning('make_unique: Converting factor to character variable')
x <- as.character(x)
}
else if(is.numeric(x)) {
if(warn_about_type_conversion) warning('make_unique: Converting numeric variable to a character vector')
x <- as.character(x)
}
deduplicated = stats::ave(x, x, FUN = function(a){
if(length(a) > 1){
suffixes <- seq_along(a)
if(wrap_in_brackets) suffixes <- paste0('(', suffixes, ')')
paste0(a, sep, suffixes)
}
else {a}
})
values_still_duplicated <- deduplicated[duplicated(deduplicated)]
if(length(stats::na.omit(values_still_duplicated)) > 0){
stop(
"make_unique failed to make vector unique.\n",
"This is because appending ' <dup_number>' to duplicate values led to",
"creation of term(s) that were in the original dataset: \n[",
paste0(values_still_duplicated, collapse = ', '),
"]\n\nPlease try again with a different argument for either `wrap_in_brackets` or `sep`"
)
}
return(deduplicated)
}