I want my data to be the "double" rather than the "integer" object category to be able to work with a specific R package, while retaining the separation of the data into distinct columns designating different sites. I have tried using the as.double() function, but this creates a string of numbers which are not separated into the columns which they have previously. Is there a way to convert to doubles to be able to work with this R package while retaining the categorisation of the data in its currents rows & columns?
To give you an example of the data I am trying to convert, and what I am trying to convert it to, here is one of the integers I have in my data:
structure(c(97L, 97L, 98L, 98L, 98L, 102L, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, NA), dim = c(6L, 3L), dimnames = list(
NULL, c("det.covs.day.day141", "det.covs.day.day142", "det.covs.day.day143"
)))
Here is a subset of that data:
det.covs.day.day141 det.covs.day.day142 det.covs.day.day143
[1,] 97 NA NA
[2,] 97 NA NA
[3,] 98 NA NA
[4,] 98 NA NA
[5,] 98 NA NA
[6,] 102 NA NA
[7,] 103 NA NA
[8,] 103 NA NA
[9,] 108 NA NA
[10,] 124 NA NA
[11,] 126 127 128
[12,] 136 137 138
[13,] 148 148 149
[14,] 188 NA NA
[15,] 204 210 NA
[16,] 204 210 NA
[17,] 204 210 NA
[18,] 204 210 NA
[19,] 204 210 NA
[20,] 204 210 NA
[21,] 204 210 NA
[22,] 204 210 NA
[23,] 204 210 NA
[24,] 204 210 NA
[25,] 204 210 NA
[26,] 204 210 NA
[27,] 204 210 NA
[28,] 204 210 NA
[29,] 204 210 NA
[30,] 205 NA NA
And here is the equivalent part of data in the tutorial for the R package:
structure(c(156, 156, 156, 156, 156, 156, 167, 167, 167, 167,
167, 167, 187, 187, 187, 187, 187, 187), dim = c(6L, 3L))
And here is a subset of the data used in the tutorial:
[,1] [,2] [,3]
[1,] 156 167 187
[2,] 156 167 187
[3,] 156 167 187
[4,] 156 167 187
[5,] 156 167 187
[6,] 156 167 187
[7,] 156 167 180
[8,] 156 167 180
[9,] 156 167 180
[10,] 156 167 180
[11,] 156 167 180
[12,] 156 167 180
[13,] 156 167 180
[14,] 156 167 180
[15,] 156 167 180
[16,] 156 167 181
[17,] 156 167 181
[18,] 156 167 181
[19,] 156 167 180
[20,] 156 167 180
[21,] 156 167 180
[22,] 156 167 180
[23,] 156 167 180
[24,] 156 167 180
[25,] 156 167 180
[26,] 156 167 180
[27,] 156 167 187
[28,] 156 167 187
[29,] 156 167 187
[30,] 156 167 187
In case it helps, the R package in question is spOccupancy https://cran.r-project.org/web/packages/spOccupancy/index.html
The two functions as.double() and as.numeric() are identical and they both convert a matrix into a vector, as you have noticed.
The way to go is to use apply() to apply the function as.numeric() (or as.double(), if you prefer) to each column of the matrix as follows:
m <- structure(c(97L, 97L, 98L, 98L, 98L, 102L, NA, NA, NA, NA, NA,
NA, NA, NA, NA, NA, NA, NA), dim = c(6L, 3L), dimnames = list(
NULL, c("det.covs.day.day141", "det.covs.day.day142", "det.covs.day.day143"
)))
m_num <- apply(m, 2, as.numeric)
m_num
## det.covs.day.day141 det.covs.day.day142 det.covs.day.day143
## [1,] 97 NA NA
## [2,] 97 NA NA
## [3,] 98 NA NA
## [4,] 98 NA NA
## [5,] 98 NA NA
## [6,] 102 NA NA
As you can see, the shape of the matrix and the column names are kept intact. The new matrix is no longer an integer:
is.integer(m_num)
## [1] FALSE
It is important to note, that integers are also numeric:
is.numeric(m)
## [1] TRUE
is.numeric(m_num)
## [1] TRUE
So, you can not use is.numeric() to distinguish between integers and doubles, but is.double() can do the job:
is.double(m)
## [1] FALSE
is.double(m_num)
## [1] TRUE
Somewhat confusingly, floating point numbers are nevertheless called "numeric" in R:
class(2.5)
## [1] "numeric"