Re: Different delimiter for 'collapse'

[Eric Powell]

Oh, well.  I rarely need multi-byte delimiters.  Thanks, anyway.

On Mon, Feb 15, 2021 at 5:26 AM Shawn Wagner <shawnw.mobile@gmail.com> wrote:

My change works for unique, too.

Using multi-byte delimiters - (Which should be done both for normal fields and for collapse/unique) isn't something it does, though. IIRC that was a bit more complicated to add.

On Sat, Feb 13, 2021 at 3:32 PM Eric Powell <powell.eric@gmail.com> wrote:

That's wonderful.  I actually like --collapse-delimiter, but for what it's worth, in Impala this would probably be called "concat".  So, maybe concat-delimeter, which would be good because it isn't specific to collapse, and as Erik pointed out 'unique' also should be considered for this new handling.

Last thing I would suggest is to allow multiple characters.  Sometimes you have an unfamiliar dataset and it is just nice to be able to set something really distinct to be safe (e.g. "@$@")

On Sat, Feb 13, 2021 at 2:10 PM Shawn Wagner <shawnw.mobile@gmail.com> wrote:

I actually have a patch to do this ready to commit when I find the time and remember to work on it... but I never was happy with the long-form name for the option I used (--collapse-delimiter). Any better suggestions?

On Sat, Feb 13, 2021 at 11:11 AM Eric Powell <powell.eric@gmail.com> wrote:

Datamash is such a wonderful piece of software and I am so happy to have discovered it.

One feature that I wish was available is to change the delimiter for the collapse operation.  My data has commas in it already so I cannot distinguish between those and the commas produced by collapse.  It would be great if there was a command-line flag allowing the user to choose the delimiter used by collapse.