In a shell script, is there a way to ONLY re-number the list, but not the contents in the list? For example:
3. cat
1. hat
2. mouse
and change it to:
1. cat
2. hat
3. mouse
Thank you.
2 Answers
Simply:
awk '$1=NR "."'
This will change the first field (first text before a blank space) to the line (record) number and a dot. Since the record number is always positive and greater than 0, the result of the assignment will always be true causing the line (record) to be printed with the first field changed.
As the first field gets changed the NF (number of fields) and the zero field ($0) get re-calculated transforming all repeated white-space to one space.
That is: lines that have multiple adjacent spaces (and tabs) are coalesced into a single space.
If multiple spaces need to be salvaged from awk processing, it becomes quite more convoluted (only GNU awk has patsplit):
awk '{
n=patsplit($0,f,/[^ \t\n]+/,s); # split the line storing whitespace.
f[1] = NR "." # only change actually needed.
for(i=0;i<=n;i++) # for all fields (and separators)
{
printf("%s%s",f[i],s[i]) # print each field and separator.
}
print("") # cause the printing of an OFS
}
' patsplitfile
cut -d. -f2- input | nl -ba -w 1 -s .
This strips the leading numbers and period from the input then pipes it to nl, which is instructed to:
-ba-- number every line-w 1-- use one column for the inserted numbers-s .-- use.as the separator between the inserted numbers and the text
On a sample input of:
3. cat
4. some words
1. hat
2. mouse
10. house
9. box
8. fox
7. here
6. there
5. anywhere
The output is:
1. cat
2. some words
3. hat
4. mouse
5. house
6. box
7. fox
8. here
9. there
10. anywhere
awk '{ $1=NR "."; print }'