File in ascending order by row

Shell Programming and Scripting
1

Hi All
I have a file like this:

ID1 ref_A 10 ref_B 30 ref_C 5
ID2 ref_F 69 ref_G 12 ref_H 5

Every ID is followed by a string(ref_X) followed by a number(every number is referred to the previous ref)
I would like to order the file like this(the column could be more, but always with the same schema- ref_X number-)

ID1 ref_C 5 ref_A 10 ref_B 30
ID2 ref_H 5 ref_G 12 ref_F 69

Hope everything is clear!
Do you have any suggestion?

Thanks

Giuliano

2

If I understand you correctly you wish to maintain the row order of the file then split each record into an ID followed by any number of ref=>value tupples and sort the fields on increasing value...

perl -ne 'chomp;($id,%r)=split;print $id;for (sort {$r{$a}<=>$r{$b}} keys %r){print " $_ $r{$_}"}print"\n";' tmp.txt
ID1 ref_C 5 ref_A 10 ref_B 30
ID2 ref_H 5 ref_G 12 ref_F 69

ETA explanation
To clarify the script above...
perl -ne ' use Perl stepping though each line of input and executing the following scriptlet
chomp; Strip off any newline character this means none of the strings will have a new line at the end of them
($id,%r)=split; tokenise the string assigning the first token to the $id variable and assign the rest of the tokens in name=> value pairs in %r
print $id; print the id
for (sort {$r{$a}<=>$r{$b}} keys %r){ step through the name=>value pairs in order of lowest value to highest value
print " $_ $r{$_}" print the name value pair out
}print"\n"; finally print a new line to replace the one we removed
' tmp.txt end of scriptlet and use tmp.txt as the input

1 Like
3

No, not everything is clear. Any attempts from your side? Show them. And, is it always the last column to be advanced, or alwas the third?

4

Thank Skrynesaver for the suggestion..I will try it immediately!
Hi RudiC! I am sorry but now I am not able to resolve an issue like that.I have start one month to study awk programming but really...it is very difficult!
However the file should be read like that:

  • column 1=ID (must be included in the output, in the first column)
  • column 2 PLUS column 3= the value of these column are associated, meaning that in column 2 there is a name (suppose "ref") and in column 3 there is a number
  • column 4 PLUS column 5=as above

etc...

What I would like is to reorder the file by maintaining in the first column the ID, and the other column (always paired) should be re-ordered in ascending order.
So

ID name_A 10 name_B 3 name_C 78 name_D 2

became

ID name_D 2 name_B 3 name_A 10 name_C 78

thanks

5

Somewhat clumsy compared to the perl oneliner above... using awk however:

awk     'function isort( A, B, n,    i, j, holdA, holdB)
                {for( i = 2 ; i <= n ; i++)
                        {holdA = A[j = i]; holdB = B[j]
                         while ( A[j-1] > holdA )
                                { j-- ; A[j+1] = A[j]; B[j+1] = B[j] }
                         A[j] = holdA; B[j] = holdB
                        }
                }

                {N=(NF-1)/2
                 for (i=1; i<=N; i++) {X=$(i*2+1); Y=$(i*2)}
                 isort (X, Y, N)
                 printf "%s ", $1
                 for (j=1; j<=N; j++) printf "%s %s ", Y[j], X[j]
                 printf "\n"
                }
        ' file
ID1 ref_C 5 ref_A 10 ref_B 30 
ID2 ref_H 5 ref_G 12 ref_F 69
1 Like