Joining lines with a common field specifically with sed?

0
votes

I've figured out how to do this with awk, but I'd like to find a way to do it with sed too. What I've seen on StackOverflow either uses awk or is extremely cryptic, doesn't work, and doesn't offer any explanation to help me make it work. I have input like this:

Hello1 aaa

Hello2 bbb

Hello1 ccc

Hello1 ddd

And need output like this:

Hello1 aaa, ccc, ddd

Hello2 bbb

If you can explain the command, that would be fantastic. I've tried reading up on sed but it's still very confusing.

2
unixawksed

2 Answers

1
votes

This might work for you (GNU sed):

sed -r '1h;1d;H;x;s/((\S+) \S+)(.*)\n\2 (\S+)$/\1,\4\3/;x;$!d;x' file

Store the first line in the hold space (HS) and then delete it. For subsequent lines, append the current line to the HS and after swapping to the HS, pattern match on like values and replace (if successful) values in the required format. Always swap back to the pattern space (PS) and then delete the PS. Finally if this is the last line swap to the HS and print it contents.

0
votes

What I've seen on StackOverflow either uses awk or is extremely cryptic, doesn't work, and doesn't offer any explanation to help me make it work.

If you can explain the command, that would be fantastic. I've tried reading up on sed but it's still very confusing.

awk '{arr[$1] = ($1 in  arr ? arr[$1] "," : "") $2 }END{ for(i in arr)print i,arr[i] }' infile

Explanation:

 awk '{
        # arr is array, index being first field
        # if array arr has key before ($1 in arr), then
        # arr[$1] contains previously stored value arr[$1] plus comma
        # and send field of line read
        # otherwise, just second field of line read

        arr[$1] = ($1 in  arr ? arr[$1] "," : "") $2 
      }
   END{ 
        # loop through array arr
        # print array key and value
        for(i in arr)print i,arr[i] 
      }
     ' infile

Test Results:

$ cat infile
Hello1 aaa
Hello2 bbb
Hello1 ccc
Hello1 ddd

$ awk '{arr[$1] = ($1 in  arr ? arr[$1] "," : "") $2 }END{ for(i in arr)print i,arr[i] }' infile
Hello1 aaa,ccc,ddd
Hello2 bbb