Re: Real number arguments for perc function

[Tim Rice]

Hi Dorian,

Thanks for your report and the kind words. The suggestion sounds valid and I've 
added it to my todo list to have a look. However, it may be difficult for me to 
find time to implement it any time soon.

If there is anything you're able to do to get started, it could help get it 
merged faster. Even just a patch to the tests to check that it's doing the 
right thing would be a start.

A patch including both test and implementation would be awesome. I could review 
this fairly quickly. A new patch might need a couple of rounds of peer review 
before it is ready for inclusion, but if you work through it with me, I'd be 
happy to end up crediting you as the author in the final git commit.

I'm also not sure when the next release will be, so even after it's 
implemented, you may need to build from source for a while.

Let me know how you would like to proceed :)

~ Tim


On Tue, Nov 05, 2024 at 05:55:42PM +1300, Dorian Garrick wrote:
> First, thanks for your development activities in datamash - it's an impressive 
> package.
>
> I was wanting to use the perc option to create credibility intervals, but it 
> seems that function only accepts an integer argument so perc:5 and perc:95 give 
> me a 90% credible interval, but I cant get a 95% credible interval as I can't 
> use perc:2.5 and perc:97.5
>
> Regards Dorian
>
> Sent with Spark