Re: uniq -z

[Egmont Koblinger]

On Fri, May 05, 2006 at 01:13:30PM +0100, Pádraig Brady wrote:

Hi,

> A work around until this gets implemented is:
> tr '\0' '\n' | uniq | tr '\n' '\0'

AFAIK the main point in preferring \0 to \n is that it can safely handle
data that may contain embedded newlines (such as filenames). Your workaround
is only valid as long as the processed data is guaranteed not to contain \n,
in which case choosing \n as delimiter throughout the whole script is most
likely simpler than \0.


bye,

Egmont