[R] Deciphering an error message

[Uwe Ligges]

Suzanne E. Blatt wrote:

> 
> 
> 
> Hello again,
> 
> Below is the code I am using which generates the error message.
> 
> Any comments greatly appreciated,
> S.

Well, my questions were:

  - 'What is a "ppp object"?'

Still not answered yet, AFAICS.

  - 'Which function has been applied and generated that error message?'

Still not answered yet, AFAICS.

You have send us a self written function, but not a reproducible example 
that helps to discover the error message. And this function does *not* 
generate the error message, but perhaps a function called within your 
function. Further on:
- You are using a function ppp() I don't know anything about.
- (At least) The line
      sp"Col <- t(trees[spColNam])
    should generate a syntax error ...


So, please use the debugging tools to find out which function produces 
the error message (e.g. via traceback()) and specify *reproducible* 
examples that are as short as possible.

We cannot help as long as you are concealing the relevant information.

Uwe


> 
> ------------------------------------
> 
> #####################################################################
> # Function to create a PPP object for all the points in a particular
> # include.  Mark the points with their species name (or a number
> # indicating it)
> #
> # Usage:
> # x <- getBigPPP(year, include, species, number=FALSE)
> #
> # Where:
> # x = ppp object with coordinates and bounds for indicated species
> #      and plots.
> # year = year(s) to include data for.  Either a single number or a
> #        vector of numbers
> # include = name(s) of plots to include data from.  Either a single
> #           character expression or a vector of characters.
> # species = name(s) of species to include data for.  Either a single
> #           species name or a vector of characters.
> # number = flag indicating whether species names or numbers are used
> #          as marks in ppp object.  Species names are default.  To
> #          use numbers, set number=TRUE.  This argument is not needed
> #          unless numbers are desired.
> #
> # Example:
> # source("spat1.R")                      # source file to install functions
> # getBigPPP(1992, "G1", "apple")         # print PPP object (summary of)
> #                                        # for apple in G1 in 1992.  Marks
> #                                        # will be species name 'apple'
> # x <- getBigPPP(1992, c("P1","P2","P3"), c("apple","hawthorn"), number=TRUE)
> #                                        # x is defined as PPP object with
> #                                        # coordinates and plot information
> #                                        # for apple and hawthorn species
> #                                        # in plots P1, P2, and P3, with
> #                                        # marks set to numbers.
> # plot(x)                                # plot locations of species, with
> #                                        # different marks for each one, and
> #                                        # plot boundaries.
> #
> getBigPPP <- function(year=NULL, include=NULL, species=NULL, number=FALSE, spColNam="Species") {
>     # message about usage
>     if(is.null(year) || is.null(include)) {
>         return("ERROR !")
>     }
>     # create species selection if it is now null
>     sp"Col <- t(trees[spColNam])
>     # print(spCol)
>     if(is.null(species)) {
>         species <- as.character(unique(spCol))
>     }
>     # create the bounds for this object
>     bounds <- plotBounds(include)
>     # selection based on include and year
>     select <- (trees$Plot %in% include) & (trees$Year %in% year) &
>         (spCol %in% species) & (!is.na(trees$Northing)) &
>         (!is.na(trees$Easting)) & (trees$Present == 1)
>     # mark with numbers or with species name
>     if(!number) {
>         return(ppp(x=trees$Easting[select], y=trees$Northing[select],
>             marks=as.factor(spCol[select]), poly=bounds))
>     }
>     specs <- unique(spCol[select])
>     marks <- apply(as.matrix(spCol[select]), 1, function(x) {
>         which(specs == x)[1]
>     })
>     return(ppp(x=trees$Easting[select], y=trees$Northing[select],
>         marks=as.factor(marks), poly=bounds))
> }
> --------------------------------------------------
> 
> 
> 
> 
> 
> 
> 
> Uwe Ligges <ligges at statistik.uni-dortmund.de> wrote:
> 
> 
>>Suzanne E. Blatt wrote:
>>
>>>Hello,
>>>
>>>I'm working in spatstat and having difficulty with the ppp objects.  I can get a ppp object for one set of data, but with the same code applied to a second data set (all I am changing is the field identifier), I get the following error message:
>>>
>>>Error in switch(w$type, rectangle={: internal error: some total scores are neither 0 nor 1
>>
>>What is a "ppp object"?
>>Which function has been applied and generated that error message?
>>
>>The error message tells you some values a ought to sum up to 0 or 1 but 
>>they don't. I (and perhaps others as well) cannot help when you don't 
>>specify some relevant information.
>>
>>Uwe Ligges
>>
>>
>>>Any thoughts on what this means and how to correct it would be most appreciated.
>>>
>>>Thanks,
>>>Suzanne
>>>