Re: The best way to convert space separated text to TSV?

[Manuel Collado]

El 11/02/2020 a las 4:42, Peng Yu escribió:
> Hi,
>
> Many programs (such as wc and ps) print results in tables with one or
> more spaces as separators. But the last column allows spaces in them.
> To process the output of wc, I came up with the following code
> (sometimes I need to manually change the display name such as
> "file1"). But it is too verbose.
>
> BEGIN {
>         OFS = "\t"
>         for(i=1;i<ARGC;++i) {
>                 fnames[i] = ARGV[i]
>         }
>         nfiles = ARGC - 1
>         delete ARGV
> }
> {
>         match($0, /^[ ]*/)
>         line = substr($0, RSTART+RLENGTH)
>         NF = 1
>         for(i=1; i<=n; ++i) {
>                 if(match(line, /[ ]+/)) {
>                         $i = substr(line, 1, RSTART-1)
>                         line = substr(line, RSTART+RLENGTH)
>                 }
>         }
>         if(NR <= nfiles) {
>                 $i = fnames[NR]
>         } else {
>                 if(line "") $i = line
>         }
>         print
> }
>
> $ awk -v n=2 -f ./wc.awk file1 <<EOF
>   a bb c
> aa  b c
> EOF
>
> $ awk -v n=3 -f ./wc.awk <<EOF
>   a bb c
> EOF
>
> What is the most succinct way to convert such kind of input to TSV
> format with gawk? Thanks.

Please try the following:

--- tabular1,awk ---
{
   nf = split($0, f, " ", s)
   offset = length(s[0])
   for (n=1; n<numcols; n++) {   # first numcols-1 columns
      offset = offset + length(f[n]) + length(s[n])
      printf("%s\t", f[n])
   }
   print substr($0, offset)      # last column, with spaces
}
--------------------

It can be invoked as:

wc xxxx | gawk -f tabular1.awk -v nomcols=4

Hope this helps.

--
Manuel Collado - http://lml.ls.fi.upm.es/~mcollado