redfit small issues is.numeric(x) is not TRUE and Error in match.fun

[chrisdane]

Hi

Two small things (dplR version 1.7.2). The last line of the redfit example

data(cana157)
yrs <- time(cana157)
x <- cana157[, 1]
redf.x <- redfit(x, nsim = 1000)

yields the error

Error in redfit(x, nsim = 1000) : is.numeric(x) is not TRUE

Instead, it must be

redf.x <- redfit(x$TTRSTD, nsim = 1000)

In addition, the very similar call

redf.x <- dplR::redfit(x$TTRSTD, nsim = 1000)

yields the error

Error in match.fun(as.list(cl)[[1]]) : 
  'as.list(dplR::redfit(x = x$TTRSTD, nsim = 1000))[[1]]' is not a function, character or symbol

Cheers, Chris

[AndyBunn]

Hi Chris. That example is set to not run because of time considerations when checking the library. But, it runs fine for me and that's what I'd expect since cana157[, 1] is cana157$TTRSTD (e.g., all(cana157[,1]==cana157$TTRSTD)). Can you confirm that this doesn't run and give me the output from sessionInfo() as well? Thx, A

library(dplR)
data(cana157)
all(cana157[,1]==cana157$TTRSTD)
x <- cana157[, 1]
redf.x <- redfit(x, nsim = 1000)

[chrisdane]

Ouh jeah it works also for me. I probably made a copy-paste mistake or whatever, sorry for this.

But the second error really is there :D

[AndyBunn]

Interesting. So this throws the error.

x <- rnorm(cumsum(100))
redf.x <- dplR::redfit(x)

But this doesn't

library(dplR)
redf.x <- redfit(x)

My initial guess is something to do with the NAMESPACE. I'll look into it when I get the chance. Thx.

[AndyBunn]

Think this is fixed in the dev version.

[japhir]

I'm still getting the same error message after installing the dev version with remote::install_github(). But I also get it after library(dplR) if I leave the function call as dplR::redfit rather than redfit.

EDIT: to make it reproducible:

# no library
data(cana157, package = "dplR")
dplR::redfit(cana157$TTRSTD)
#> Error in match.fun(as.list(cl)[[1]]) : 
#>  'as.list(dplR::redfit(x = cana157$TTRSTD))[[1]]' is not a function, character or symbol

library(dplR)
#> This is dplR version 1.7.7.
#> dplR is part of openDendro https://opendendro.org.
#> New users can visit https://opendendro.github.io/dplR-workshop/ to get started.

dplR::redfit(cana157$TTRSTD)
#> Error in match.fun(as.list(cl)[[1]]) : 
#>  'as.list(dplR::redfit(x = cana157$TTRSTD))[[1]]' is not a function, character or symbol

redfit(cana157$TTRSTD)
#> redfit() in dplR version 1.7.7
#> 
#> Input:
#> ------
#> ofac = 4
#> hifac = 1
#> n50 = 3
#> iwin = 2
#> nsim = 1000
#> 
#> Initial values:
#> ---------------
#> Data variance (from data spectrum) = 0.02679094
#> Avg. dt = 1
#> 
#> Results:
#> --------
#> Avg. autocorr. coeff., rho = 0.5233948
#> Avg. tau = 1.544594
#> Degrees of freedom = 5.785714
#> 6-dB Bandwidth = 0.00862069
#> Critical false-alarm level (Thomson, 1990) = 99.56897
#>    ==> corresponding scaling factor for red noise = 3.203169
#> 
#> Equality of theoretical and data spectrum: Runs test
#> ----------------------------------------------------
#> Test requires iwin = 0
#> Test requires ofac = 1
#> Test requires n50 = 1