Extract numeric part of strings of mixed numbers and characters in R

Question

I have a lot of strings, and each of which tends to have the following format: Ab_Cd-001234.txt I want to replace it with 001234. How can I achieve it in R?

Answer 1

The stringr package has lots of handy shortcuts for this kind of work:

# input data following @agstudy
data <-  c('Ab_Cd-001234.txt','Ab_Cd-001234.txt')

# load library
library(stringr)

# prepare regular expression
regexp <- "[[:digit:]]+"

# process string
str_extract(data, regexp)

Which gives the desired result:

  [1] "001234" "001234"

To explain the regexp a little:

[[:digit:]] is any number 0 to 9

+ means the preceding item (in this case, a digit) will be matched one or more times

This page is also very useful for this kind of string processing: http://en.wikibooks.org/wiki/R_Programming/Text_Processing

Answer 2

Using gsub or sub you can do this :

 gsub('.*-([0-9]+).*','\\1','Ab_Cd-001234.txt')
"001234"

you can use regexpr with regmatches

m <- gregexpr('[0-9]+','Ab_Cd-001234.txt')
regmatches('Ab_Cd-001234.txt',m)
"001234"

EDIT the 2 methods are vectorized and works for a vector of strings.

x <- c('Ab_Cd-001234.txt','Ab_Cd-001234.txt')
sub('.*-([0-9]+).*','\\1',x)
"001234" "001234"

 m <- gregexpr('[0-9]+',x)
> regmatches(x,m)
[[1]]
[1] "001234"

[[2]]
[1] "001234"

Answer 3

You could use genXtract from the qdap package. This takes a left character string and a right character string and extracts the elements between.

library(qdap)
genXtract("Ab_Cd-001234.txt", "-", ".txt")

Though I much prefer agstudy's answer.

EDIT Extending answer to match agstudy's:

x <- c('Ab_Cd-001234.txt','Ab_Cd-001234.txt')
genXtract(x, "-", ".txt")

# $`-  :  .txt1`
# [1] "001234"
# 
# $`-  :  .txt2`
# [1] "001234"

Answer 4

gsub Remove prefix and suffix:

gsub(".*-|\\.txt$", "", x)

tools package Use file_path_sans_ext from tools to remove extension and then use sub to remove prefix:

library(tools)
sub(".*-", "", file_path_sans_ext(x))

strapplyc Extract the digits after - and before dot. See gsubfn home page for more info:

library(gsubfn)
strapplyc(x, "-(\\d+)\\.", simplify = TRUE)

Note that if it were desired to return a numeric we could use strapply rather than strapplyc like this:

strapply(x, "-(\\d+)\\.", as.numeric, simplify = TRUE)

Answer 5

I'm adding this answer because it works regardless of what non-numeric characters you have in the strings you want to clean up, and because OP said that the string tends to follow the format "Ab_Cd-001234.txt", which I take to mean allows for variation.

Note that this answer takes all numeric characters from the string and keeps them together, so if the string were "4_Ab_Cd_001234.txt", your result would be "4001234".

If you're wanting to point your solution at a column in a dataframe you've got,

df$clean_column<-gsub("[^0-9]", "", df$dirty_column)

This is very similar to the answer here: https://stackoverflow.com/a/52729957/9731173.

Essentially what you are doing with my solution is replacing any non-numeric character with "", while the answer I've linked to replaces any character that is not numeric, - or .