Best way to split input lines based on fixed field widths?

[ollystephens]

I recently discovered that gnu awk has a way of dealing with fixed-width fields with no distinctive field separator.

https://www.gnu.org/software/gawk/manual/html_node/Constant-Size.html

As I've just encountered output that fits this category, and as I like to do as much as I can with Miller, I was wondering on best approach.

For now, I'm filtering the output through gawk and into Miller in order to add a structure Miller can parse, but I'd rather do it all in one tool.

[aborruso]

A related issue #224

[johnkerl]

@ollystephens @aborruso indeed, Miller does not currently have this, but it would be nice.

One use-case is for 'recovering' pprint output which can be lossy with respect to delimiters. For example:

$ cat space.csv
A,B,C
s,t u,v

$ mlr -i csv -o json cat space.csv
[
{
  "A": "s",
  "B": "t u",
  "C": "v"
}
]

$ mlr -i csv -o pprint cat space.csv
A B   C
s t u v

$ mlr -i csv -o pprint cat space.csv | mlr -i pprint -o json cat
mlr: mlr: CSV header/data length mismatch 3 != 4 at filename (stdin) line 2.