How to validate whether all csv files have the same first line?

I have a directory containing many csv files. I want a shell script to check whether the first line of each file is the same. For example, these files have the same header, so the check should return True.

❯ cat file1.csv
column1,column2,column3
3,1,3
4,3,9
❯ cat file2.csv
column1,column2,column3
5,4,1
1,8,2

I thought this would be an already asked question, but I haven't seen it on Stackoverflow.

This question is similar, which checks whether all lines in one file are the same: https://unix.stackexchange.com/questions/533915/check-if-all-lines-in-a-file-are-same.

What I've tried:

• echo "$(ls -AU | head -1)" gets me the first line of one file
• I thought about trying to assert that all files have this value as a first line (preferably as a concise pipe rather than for loop), but couldn't figure out how to do this
• I've tried to use the answer here https://unix.stackexchange.com/a/533917, which uses uniq and wc, but both commands seem specific to iterating through lines of a single file (rather than iterating over a generic list output)

With GNU sed and bash:

rows=$(sed -s '1!d' *.csv | sort -u | wc -l)
if [[ "$rows" -eq 1 ]]; then echo "true"; else echo "false"; fi

sed -s '1!d' *.csv outputs to stdout first line of all files in current directory with suffix .csv.

It is helpful to know WHICH file is different.

Given:

head -n 2 *.csv
==> file1.csv <==
column1,column2,column3
3,1,3

==> file2.csv <==
column1,column2,column3
5,4,1

==> file3.csv <==
column_diff,column2,column3
5,4,1

==> file4.csv <==
column1,column2,column3
5,4,1

You can use this Ruby to determine the offender:

ruby -e 'keys=ARGV.each_with_object(Hash.new {|h,k| h[k] = []}){|fn,h| 
	h[File.open(fn).readline.chomp]<<fn
}
keys.each{|k,v| puts "#{k}\n\t#{v.join("\n\t")}"}
' *.csv

Prints:

column1,column2,column3
	file1.csv
	file2.csv
	file4.csv
column_diff,column2,column3
	file3.csv

If you want to be able to test from Bash the result:

ruby -e 'keys=ARGV.each_with_object(Hash.new {|h,k| h[k] = []}){|fn,h| 
	h[File.open(fn).readline.chomp]<<fn
}
if keys.length>1 then
	keys.each{|k,v| puts "#{k}\n\t#{v.join("\n\t")}"}
	exit(false)
else
	puts "All files equal"
end
' *.csv

Then just test the exit code.

Using GNU head (for the -q option), this one-liner should do the trick:

if [ -z "$(head -q -n1 *.csv | uniq -u)" ]; then echo true; else echo false; fi

The -q option of GNU head will suppress printing of filenames. uniq -u will give no output only if all lines are the same. sort isn't necessary for this task.

Using any awk (untested):

awk '
    FNR == 1 {
        if ( $0 != prev ) {
            status = 1
            exit
        }
        prev = $0
        nextfile
    }
    END {
        print ( status ? "false" : "true" )
        exit status
    }
' *.csv

It'll run faster if your awk supports nextfile but it'll work either way.

I'm assuming above that you want your script both print "true"/"false" and to exit with a success/fail status.