AWK: 2 columns 2 files display where second column has unique data - awk

Question

I need the first column to check if it doesn't match the first column on the second file. Though, if the second column matches the second column on the second file, to display this data with awk on Linux.
I want awk to detect the change with both the first and second column of the first file with the second file.
file1.txt
sdsdjs ./file.txt
sdsksp ./example.txt
jsdjsk ./number.txt
dfkdfk ./ok.txt
file2.txt
sdsdks ./file.txt <-- different
sdsksd ./example.txt <-- different
jsdjsk ./number.txt <-- same
dfkdfa ./ok.txt <-- different
Expected output:
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
Notice how in the second file there may be lines missing and not the same as the first.
As seen above, how can awk display results only where the second column is unique and does not match the first column?

Answer 1

Something like this might work for you:
awk 'FNR == NR { f[FNR"_"$2] = $1; next }
f[FNR"_"$2] && f[FNR"_"$2] != $1' file1.txt file2.txt
Breakdown:
FNR == NR { } # Run on first file as FNR is record number for the file, while NR is the global record number
f[FNR"_"$2] = $1; # Store first column under the name of FNR followed by an underbar followed by the second column
next # read next record and redo
f[FNR"_"$2] && f[FNR"_"$2] != $1 # If the first column doesn't match while the second does, then print the line
A simpler approach which will ignore the second column is:
awk 'FNR == NR { f[FNR"_"$1] = 1; next }
!f[FNR"_"$1]' file1.txt file2.txt

Answer 2

If the records don't have to be in the respective position in the files ie. we compare matching second column strings, this should be enough:
$ awk '{if($2 in a){if($1!=a[$2])print $2}else a[$2]=$1}' file1 file2
Output:
file.txt
In pretty print:
$ awk '{
if($2 in a) { # if $2 match processing
if($1!=a[$2]) # and $1 don t
print $2 # output
} else # else
a[$2]=$1 # store
}' file1 file2
Updated:
$ awk '{if($2 in a){if($1!=a[$2])print $1,$2}else a[$2]=$1}' file1 file2
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
Basically changed the print $2 to print $1,$2.

Answer 3

The way your question is worded is very confusing but after reading it several times and looking at your posted expected output I THINK you're just trying to say you want the lines from file2 that don't appear in file1. If so that's just:
$ awk 'NR==FNR{a[$0];next} !($0 in a)' file1 file2
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
If your real data has more fields than shown in your sample input but you only want the first 2 fields considered for the comparison then fix your question to show a more truly representative example but the solution would be:
$ awk 'NR==FNR{a[$1,$2];next} !(($1,$2) in a)' file1 file2
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
if that's not it then please edit your question to clarify what it is you're trying to do and include an example where the above doesn't produce the expected output.

Answer 4

I understand the original problem in the following way:
Two files, file1 and file2 contain a set of key-value pairs.
The key is the filename, the value is the string in the first column
If a matching key is found between file1 and file2 but the value is different, print the matching line of file2
You do not really need advanced awk for this task, it can easily be achieved with a simple pipeline of awk and grep.
$ awk '{print $NF}' file2.txt | grep -wFf - file1.txt | grep -vwFf - file2.txt
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
Here, the first grep will select the lines from file1.txt which do have the same key (filename). The second grep will try to search the full matching lines from file1 in file2, but it will print the failures. Be aware that in this case, the lines need to be completely identical.
If you just want to use awk, then the above logic is achieved with the solution presented by Ed Morton. No need to repeat it here.

Answer 5

I think this is what you're looking for
$ awk 'NR==FNR{a[$2]=$1; next} a[$2]!=$1' file1 file2
sdsdks ./file.txt
sdsksd ./example.txt
dfkdfa ./ok.txt
print the records from file2 where field1 values are different for the same field2 value. This script assumes that field2 values are unique within each file, so that it can be used as keys. Since the content looks like file paths, this is a valid assumption. Otherwise, you need to match the records perhaps with the corresponding line numbers.

Answer 6

In case you are looking for a more straightforward line-based diff based on the the first field on a line being different.
awk 'NR==FNR { a[NR] = $1; next } a[FNR]!=$1' file1 file2