extract values from string with random pattern

April 25, 2024

How can I extract 1.10E+93, 1012055, 1018598,AOL,WDF985 from the strings vector as below using pattern? I want to extract two times 1.10E+93 and 1012055 and one time other values i.e., extract one value for each string.

string from which to extract values

strings <- c('/ccr/1.10E+93_ccrdt/indices/1.10E+93_ccr_ann_123.csv',
             '/ccr/1.10E+93_ccrdt/indices/1.10E+93_obsrst_ann.csv',
             '/ccr/1012055_obsrt/indices/1012055_obsrrt.csv',
             '/ccr/1012055_obsrt/indices/1012055_ccr_ann.csv',
             '/ccr/1018598_obsrt/indices/1018598_obsrrt.csv',
             '/ccr/AOL_obsrt/indices/AOL_rrst.csv',
             '/ccr/WDF985_obsrt/indices/WDF985_rrst.csv')

>Solution :

There are lot of ways to do this (Copilot or ChatGPT might give you a good answer.) Maybe you want

strings |> 
   ## remove everything up to the last slash
   stringr::str_remove("^.*/") |> 
   ## remove underscore and everything after it
   stringr::str_remove("_.*$")
[1] "1.10E+93" "1.10E+93" "1012055"  "1012055"  "1018598"  "AOL"      "WDF985"

stringr::str_extract(strings, "/([^/_]*)_", group = TRUE)

Or (base R)

gsub("^.*/([^_]*)_.*$", "\\1", strings)

You can also use stringr::str_extract(). Lookbehind/lookahead components of the regular expression are useful for requirements of the form "extract all characters between [A] and [B] but don’t include [A] or [B] in the result"