Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems. It only takes a minute to sign up.
Sign up to join this community
I know how to select a field from a line using the cut command. For instance, given the following data:
a,b,c,d,e
f,g,h,i,j
k,l,m,n,o
This command:
cut -d, -f2 # returns the second field of the input line
Returns:
b
g
l
My question: How can I select the second field counting from the end? In the previous example, the result would be:
d
i
n
Reverse the input before and after cut with rev:
<infile rev | cut -d, -f2 | rev
Output:
d
i
n
cut can't take negative field indices (like Python).
Apr 9, 2018 at 4:55
Try doing this with awk :
awk -F, '{print $(NF-1)}' file.txt
Or using perl :
perl -F, -lane 'print $F[-2]' file.txt
Or using ruby (thanks manatwork) :
ruby -F, -lane 'print $F[-2]' file.txt
Or using bash (thanks manatwork) :
while IFS=, read -ra d; do echo "${d[-2]}"; done < file.txt
Or using python :
cat file.txt |
python -c $'import sys\nfor line in sys.stdin:\tprint(line.split(",")[-2])'
bash not needs fixed column count for this: while IFS=, read -ra d; do echo "${d[-2]}"; done < file.txt.
Feb 13, 2013 at 15:03
- or (depending on the environment, shell, or how the shell was compiled), may contain backslash characters, then echo is not an option. Why do you need to concatenate file.txt with nothing before feeding it to python!?. You need read -A instead of read -a in ksh93 and zsh. Negative subscripts work in zsh but only in recent versions of ksh93 and bash. In older versions, you can use ${d: -2:1}
Feb 13, 2013 at 16:01
${d[@]: -2:1} in your last sentence.
Feb 15, 2013 at 8:17
Using sed:
sed -E 's/^([^,]*,)*([^,]*)(,[^,]*){1}$/\2/' infile
Output:
d
i
n
([^,]*,)* matches any number of non-comma charecter followed by a comma, i.e. any number of columns.([^,]*) matches a column.(,[^,]*){1} matches one column at the end, if you change the quantifier {1} to {2} it matches the third column from the end etc.