Fix rhs/data issue in amova

[zkamvar]

When running amova using a data frame, it would error because of line 18, where it was trying to retrieve an object that didn't exist. This is bypassed if the data is present in a data.frame:

data(microbov)
set.seed(20150409)
mic20 <- microbov[sample(nInd(microbov), 20)]
microdf <- data.frame(other(mic20))
d <- dist(tab(mic20))
amova(d ~ spe/breed, data = microdf, nperm = 0)

[zkamvar]

Any word on this? Good? Bad?

[emmanuelparadis]

Wed, 29 Apr 2015 08:45:42 -0700 "Zhian N. Kamvar"
notifications@github.com:

Any word on this? Good? Bad?

Thanks for the reminder. I'll have a look at it very soon.

[emmanuelparadis]

There seems to be a problem when the rhs is not a factor. I get an error:

> res <- fixed.amova(d ~ g, DF, nperm = 100)
Error in FUN(X[[i]], ...) : 'bin' must be numeric or a factor
> str(DF$g)
 chr [1:15] "A" "A" "A" "A" "A" "A" "A" ...

But not if the rhs is in the global env.:

> res <- fixed.amova(d ~ g, nperm = 100)
Warning message:
In fixed.amova(d ~ g, nperm = 100) :
  elements in the rhs of the formula are not all factors
> str(g)
 chr [1:15] "A" "A" "A" "A" "A" "A" "A" ...

It's just a warning because of StAMMP (I guess you've seen the comment in the source).

I think there's a possibility to use the envir option in get even with a data.frame:

> get("g", envir = as.environment(DF))
 [1] "A" "A" "A" "A" "A" "A" "A" "B" "B" "B" "B" "B" "B" "B"
[15] "B"

This'd solve the initial bug.

[zkamvar]

Good points. I think I have addressed this in my latest commit. I created a new variable to contain the environment where the data lives and placed an explicit "get" call in the control structure checking for factors.

[emmanuelparadis]

Perfect! I've merged and updated the files. Many thanks for this.