Re: [bug-gawk] Count empty line as a column with one empty field

[arnold]

Peng Yu <address@hidden> wrote:

> Hi,
>
> I see this.
>
> $ printf '%s\n' a '' b | awk -e '{print NF}'
> 1
> 0
> 1
>
> Is it possible to consider the empty line to contain one empty field
> instead of no fields?
>
> [ ..... ]
>
> I am not saying it is a bug, either.
>
> A row with an empty field, instead of an empty row, is just another
> way to interpret in the input. Both are fine depending on the context.
> In certain cases, the former may be preferred over the latter.

There isn't any way to do this with the language. But as Davide Brini
suggests:

> $ printf '%s\n' a '' b | awk '!NF{$1=""} {print NF}'
> 1
> 1
> 1

Arnold