extract different lines from files using Bash

问题:

I have two files and I use the "comm -23 file1 file2" command to extract the lines that are different from a file to another.

I would also need something that extracts the different lines but also preserves the string "line_$NR". Example: file1:

line_1: This is line0
line_2: This is line1
line_3: This is line2
line_4: This is line3

file2:

line_1: This is line1
line_2: This is line2
line_3: This is line3

I need this output: differences file1 file2:

line_1: This is line0.

In conclusion I need to extract the differences as if the file has not line_$NR at the beginning but when I print the result I need to also print line_$NR.

回答1:

Try using awk

awk -F: 'NR==FNR {a[$2]; next} !($2 in a)' file2 file1

Output:

line_1: This is line0

Short Description

awk -F: '             # Set filed separator as ':'. $1 contains line_<n> and $2 contains 'This is line_<m>'
    NR==FNR {         # If Number of records equal to relative number of records, i.e. first file is being parsed
        a[$2];        # store $2 as a key in associative array 'a'
        next          # Don't process further. Go to next record.
    } 
    !($2 in a)        # Print a line if $2 of that line is not a key of array 'a'
' file2 file1

---

Additional Requirement (In comment)

And if I have multiple ":" in a line : "line_1: This :is: line0" doesn't work. How can I only take the line_x

In that case, try following (GNU awk only)

awk -F'line_[0-9]+:' 'NR==FNR {a[$2]; next} !($2 in a)' file2 file1

回答2:

this awk line is longer, however it would work no matter where the differences were located:

awk 'NR==FNR{a[$NF]=$0;next}a[$NF]{a[$NF]=0;next}7;END{for(x in a)if(a[x])print a[x]}' file1 file2

test:

kent$  head f*
==> f1 <==
line_1: This is line0
line_2: This is line1
line_3: This is line2
line_4: This is line3

==> f2 <==
line_1: This is line1
line_2: This is line2
line_3: This is line3

#test f1 f2
kent$  awk 'NR==FNR{a[$NF]=$0;next}a[$NF]{a[$NF]=0;next}7;END{for(x in a)if(a[x])print a[x]}' f1 f2
line_1: This is line0

#test f2 f1:    
kent$  awk 'NR==FNR{a[$NF]=$0;next}a[$NF]{a[$NF]=0;next}7;END{for(x in a)if(a[x])print a[x]}' f2 f1
line_1: This is line0