Re: st: rounding the minimum of a negative number

[annoporci <annoporci@gmail.com>]

> The best starting point is now William Gould's blog entry
>
> http://blog.stata.com/2012/04/02/the-penultimate-guide-to-precision/

Thanks a lot for the reference Nick, I'll look it up.


> 1. At machine level Stata's calculations are necessarily in binary.
> Stata and most modern software go to enormous lengths to make it seem
> otherwise, but the fact is inescapable, ineluctable and inevitable.

Sure, I can't argue with that, but I would have expected Stata to keep a  
few digits of zeros stored in memory (say 6 zeros), instead of making up  
its own digits.


> Somewhere in StataCorp's files a suggestion of mine that a second  
> argumentbe allowed for these functions

That would be a great addition!


> The best you can do, I believe, is indeed something like
>
> gen double myresult = floor(100 * mydata) / 100

Got it, thanks Nick.


> You mention r(min) in your example; presumably that arises from some
> previous calculation. It is not a general solution.

Sorry, I was imprecise, the r(min) comes after -sum-, at the end of the  
code given at the bottom of my message.


> P.S. Your last example brings in a different point. `r(min)' is
> related to r(min) as local macros are to scalars. Macros hold less
> precision than scalars. For maximum precision, avoid the numeric --
> string -- numeric cycling required by putting a numeric result into a
> local macro and then taking it out again. (This won't bite with small
> or moderate integers.)

I see, I hadn't noticed that macros hold less precision than scalars.

Great tips, thanks a lot for your help Nick.

--
Patrick Toche.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/