Replace each successive instance of a pattern in one file with each successive item in another file - awk

Question

Using awk or sed how can I loop through each instance of a pattern in one file and replace it with the contents of another file. For example:
File1.txt:
123 zone
123 patch
XXX family
456 zone
456 patch
XXX family
789 zone
789 patch
XXX family
File2.txt:
123
456
789
I know how to replace all instances of a pattern to the same number, i.e. -
{a = 0;}
($2 == "family") {printf("%f %s\n",$1=123,$2); a=1;}
{a = 0;}
but I do not now how to change each successive instance of a pattern to a different number - supplied in a list of numbers.
I want to replace the first instance of XXX family in File 1 with the first number in File 2, replace the second instance of XXX family in File 1 with the second number in File 2, etc... so I end up with:
New_file.txt
123 zone
123 patch
123 family
456 zone
456 patch
456 family
789 zone
789 patch
789 family

Answer 1

This might work for you (GNU sed):
sed -E '1{x;s/.*/cat file2.txt/e;s/ //g;x}
/XXX/{G;s/XXX([^\n]*)\n([^\n]*)/\2\1/;P;s/[^\n]*\n//;h;d}' file1.txt
Prime the hold space with the contents of file2.txt (remove any trailing spaces on each line).
Match XXX and if so append the hold space, use pattern matching to replace XXX with the contents of the first line of the hold space. Print the result and then remove the first line and restore the remainder of the hold space, ready for the next replacement.

Answer 2

----------- Edit -----------------
From Glenn Jackman's very helpful and correct comment
awk 'FNR==NR {chk[FNR]=$1; next}
$1 == "XXX" {
$1 = chk[++j]
}
{print}' f2.txt f1.txt
Here is a verbal description of what is happening
(awk can read any number of files from it's command line arguments (limited by the OS cmd-line size) )
We pass in the fix list file (f2.txt) first, so it can be captured into an array. Fortunately your data is in order so we use chk[1]-chk[3] as the keys to those values. next skips any further code in the file, but reading the next record from f2 until all data is stored in the chk array. No records have been printed yet.
The rest of the code processes records from the f1.txt file only (which is the 2nd file in the list). It checks that the record's first field is XXX, using $1 == "XXX". If that test is true, we cycle thru the list stored in the chk[] array, and replace the XXX field with the next value from chk[++j]. We are using the j variable as a counter to index the elsments of the chk[] array.
As all records need to be printed, including the now "fixed" record, we use the print command.
(If you need to learn about ++ for variables, you'll do best to consult a programming book, as it is often the topic for a complete chapter.)
------------- Original answer ------------------
awk 'FNR==NR{
chk[++i]=$1;next
}
{
if (FNR!=NR && $1 ~ /^XXX/){
$1=chk[++j]
print $0
}
else {
print $0
}
}' f2.txt f1.txt
output
123 zone
123 patch
123 family
456 zone
456 patch
456 family
789 zone
789 patch
789 family
awk can read any number of files from it's command line arguments (limited by the OS cmd-line size)
We pass in the fix list file (f2.txt) first, so it can be captured into an array. Fortunately your data is in order so we use chk[1]-chk[3] as the keys to those values. next skips any further processing to get the next record from f2 until all data is stored in the chk array.
The second block only processes records from the f1.txt (which is the 2nd file in the list) . This is done with the check FNR!=NR, which you can read all about in various awk books (this is a common pattern for multi-file processing where 2 "types" of data are being processed).
AHD we check that the record begins XXX, using a regexp /^XXX/. ^ means "anchored to the beginning of the line".
We cycle thru the list stored in the chk[] array, and replace the XXX field with the next value from chk[++j].
With final else, we print any records that haven't been modified.