Conditional gsub action - if-statement

Question

According to this answer, I am trying to reproduce a conditional statement where, in the event of a match, a substitusion occurs (it matches dates). If no match happens, the line is printed as it is.
#!/bin/bash
cleaner(){
./date_remove.awk $1
}
cleaner $1 > "out"
where 'date_remove.awk' is
#! /usr/bin/awk -f
date = /(^|[^[:alpha:]])[[:digit:]]{2}[[:space:]]{1,}[[:alpha:]]{3,8}[[:space:]]{1,}[[:digit:]]{4}([^[:alpha:]]|$)/ {gsub(date, "")} !date {print}
At this point the substitution does not happens. 'gsub' should return only the matched phrases, but it does not return anything, actually. Just unmatched phrases are printed correctly. At this point, I am pretty sure is a problem of syntax, but I cannot figure out where.
Input:
ci sono 4444444444444Quattro mele
sentiamoci il 16 Ottobre 2018
deciIIIIIIdiamo il 17 ottabre 2017
Manipolo di eroi 55555555555
17 mele
18 ott 2020 llllllLLLLLLLLLLLL
una mela e mezza
2 mAAAeleA
0000 asd a0 0 ad000
Actual output:
ci sono 4444444444444Quattro mele
Manipolo di eroi 55555555555
17 mele
una mela e mezza
2 mAAAeleA
0000 asd a0 0 ad000
Expected output:
ci sono 4444444444444Quattro mele
sentiamoci il
deciIIIIIIdiamo il
Manipolo di eroi 55555555555
17 mele
llllllLLLLLLLLLLLL
una mela e mezza
2 mAAAeleA
0000 asd a0 0 ad000

Answer 1

It is not quite correct, gsub() does not return the matched phrases on its own. It just returns the count of substitutions made. Your problem is dealing with how to store the matching group for subsequent string replacement.
The problem with your attempt is the regexp matched within /../ is not stored explicitly, you need to make it be stored by using match() or index() and use that in the replacement part,
awk '
match($0, /(^|[^[:alpha:]])[[:digit:]]{2}[[:space:]]{1,}[[:alpha:]]{3,8}[[:space:]]{1,}[[:digit:]]{4}([^[:alpha:]]|$)/) {
str=substr($0, RSTART, RLENGTH); sub(str," ",$0 );
}1' file
The example above would replace the captured group i.e. your date strings below and replace them with a single white space.
16 Ottobre 2018
17 ottabre 2017
18 ott 2020
One could use sub() or gsub() depending on the number of occurrences of the regex in the line. Applying the command above would remove the those date strings from the file and produce a result as below.
ci sono 4444444444444Quattro mele
sentiamoci il
deciIIIIIIdiamo il
Manipolo di eroi 55555555555
17 mele
llllllLLLLLLLLLLLL
una mela e mezza
2 mAAAeleA
0000 asd a0 0 ad000
Notice the {..}1 after we do the string replace. It is needed to reconstruct the line after the appropriate replacements are done.
Putting it in awk script it would look like
#!/usr/bin/awk -f
match($0, /(^|[^[:alpha:]])[[:digit:]]{2}[[:space:]]{1,}[[:alpha:]]{3,8}[[:space:]]{1,}[[:digit:]]{4}([^[:alpha:]]|$)/) {
str=substr($0, RSTART, RLENGTH)
sub(str," ",$0 )
}1