R help - Feb 2015 - Saving Mean Relative Difference from all.equal()

Hello,

Does anyone know how to save the numeric value of the "mean relative
difference" when using the all.equal() command?

For example this:

all.equal(cov2cor(ITEMCOV),cor(item.data))

Gives:

[1] "Attributes: < Length mismatch: comparison on first 1 components
>"
[2] "Mean relative difference: 0.01523708"   

I'd like to save the value 0.01523708 in a numeric format.

Thanks,




--
View this message in context:
http://r.789695.n4.nabble.com/Saving-Mean-Relative-Difference-from-all-equal-tp4703905.html
Sent from the R help mailing list archive at Nabble.com.

I think I have one solution. Not very pretty though. Relies on the text not
changing at all.

as.numeric(gsub("Mean relative difference: ", "",
all.equal(cov2cor(ITEMCOV),cor(item.data))[2]))

Is there a better way?



--
View this message in context:
http://r.789695.n4.nabble.com/Saving-Mean-Relative-Difference-from-all-equal-tp4703905p4703908.html
Sent from the R help mailing list archive at Nabble.com.

> On Feb 26, 2015, at 2:02 PM, Scott Colwell <scolwell at uoguelph.ca>
wrote:
> 
> I think I have one solution. Not very pretty though. Relies on the text not
> changing at all.
> 
> as.numeric(gsub("Mean relative difference: ", "",
> all.equal(cov2cor(ITEMCOV),cor(item.data))[2]))
> 
> Is there a better way?
> 
`all.equal` is a generic function and it appears you are interested in
`all.equal.numeric`. Not sure if it?s ?better? but you could fairly easily hack
the all.equal.numeric function to replace this code (near the end of the
function code) :

if (is.na(xy) || xy > tolerance) 
        msg <- c(msg, paste("Mean", what, "difference:",
format(xy)))

,,, with:

if (is.na(xy) || xy > tolerance) 
        msg <-  xy

(probably best to make it have a different name and not use 'all.equal'
to name it.)

When it did that I got this as a result:
> all.EQ( seq(1,2, by=0.1)*10, 10:20)
[1] TRUE
> all.EQ( seq(1,2, by=0.11)*10, 10:20)
[1] "Numeric: lengths (10, 11) differ"
> all.EQ(seq(1,2, by=0.11)*10, 10:19)
[1] 0.03225806

? 


David Winsemius, MD
Alameda, CA, USA