I have a long string. The part is
x <- "Text1 q10_1 text2 q17 text3 q22_5 ..."
How can I subtract 1 from each number after "q" letter to obtain the following?
y <- "Text1 q9_1 text2 q16 text3 q21_5 ..."
I can extract all my numbers from x:
numbers <- stringr::str_extract_all(x, "(?<=q)\\d+")
numbers <- as.integer(numbers[[1]]) - 1
But how can I update x with these new numbers?
The following is not working
stringr::str_replace_all(x, "(?<=q)\\d+", as.character(numbers))
I learned today that stringr::str_replace_all will take a function:
stringr::str_replace_all(
x,
"(?<=q)\\d+",
\(x) as.character(as.integer(x) - 1)
)
We can use gregexpr and regmatches for this:
x <- "Text1 q10_1 text2 q17 text3 q22_5 ..."
gre <- gregexpr("(?<=q)[0-9]+", x, perl = TRUE)
regmatches(x, gre)
# [[1]]
# [1] "10" "17" "22"
regmatches(x, gre) <- lapply(regmatches(x, gre), function(z) as.integer(z) - 1L)
x
# [1] "Text1 q9_1 text2 q16 text3 q21_5 ..."
Here is an alternative approach using gsubfn function from gsubfn package:
gsubfn matches "q" followed by one or more digits.~ paste0("q", as.numeric(x) - 1) which converts the captured digits into numeric values and subtracts 1, and then concatenates it back.#install.packages("gsubfn")
library(gsubfn)
gsubfn("q(\\d+)", ~ paste0("q", as.numeric(x) - 1), x)
"Text1 q9_1 text2 q16 text3 q21_5 ..."
Here is a rather convoluted solution that should still work:
library(dplyr)
library(stringr)
library(magrittr)
library(tidyr)
x <- "Text1 q10_1 text2 q17 text3 q22_5 ..."
x <- data.frame(x, stringsAsFactors = FALSE) #Convert to data.frame
x %<>%
separate_longer_delim(cols = everything(), " ") %>% #Split the string into separate rows.
mutate(s = as.numeric(str_extract(x, "(?<=q)[0-9]+"))-1) %>% #Extract the value and subtract 1 from it.
mutate(x1 = str_replace(x, "(?<=q)[0-9]+", as.character(s))) %>% #Replace the value with this new value.
select(x1) %>% #Retain only data necessary to produce the output.
mutate(x1 = paste0(x1, collapse = " ")) %>% #Collapse this into a string.
distinct(x1) #Discard superfluous copies of the string.
x <- x$x1
x
# [1] "Text1 q9_1 text2 q16 text3 q21_5 ..."
here is a python version that uses logic in place of regular expressions. It can probably be cleaned up a bit...
x = "Text1 q10_1 text2 q17 text3 q22_5 ..."
list_x = x.split()
list_y = []
for y in list_x:
if y[0] == 'q':
z = y.find('_')
if z == -1:
z = None
end = ''
else: end = y[z:]
r = int(y[1:z])
r = r-1
r_string = 'q' + str(r) + str(end)
else:
r_string = y
list_y.append(r_string)
final = ' '.join(list_y)
print(final)
The result is:
Text1 q9_1 text2 q16 text3 q21_5 ...
Having shown the above, regex is a more suitable method.
