---
title: "Common Issues : F Matrix"
output:
   rmarkdown::html_vignette:
     tabset: true
bibliography: references.bib
vignette: >
  %\VignetteIndexEntry{Common Issues : F Matrix}
  %\VignetteEncoding{UTF-8}
  %\VignetteEngine{knitr::rmarkdown}
editor_options: 
  chunk_output_type: console
---

# non-conformable arguments
> Error in Fn %*% t(S) : non-conformable arguments

This error can mean there is some disagreement between the S and F matricies.

# non-numeric argument to binary operator
> non-numeric argument to binary operator

This error can mean that the F matrix has extra columns that should not be there.
* [ref](https://github.com/phytoclass/phytoclass/issues/10)

# x needs two+ dimensions
> 'x' must be an array of at least two dimension

https://github.com/phytoclass/phytoclass/issues/11

# subscript out of bounds:
> Error in x[subset & !is.na(subset), vars, drop = drop] : subscript out of bounds

https://github.com/phytoclass/phytoclass/issues/12

This error can happen when the sample names are included in the S matrix.
Current version of phytoclass does not support sample names.

# 'x' must be numeric
> Error in rowSums(Fmat) : 'x' must be numeric

This can mean that the F matrix has a numeric index (1,2,3,etc) column that must be removed.
This is sometimes introduced by `read.csv` or `readRDS`.
A workaround for this:

```{R}
#| code-summary: clean up the F matrix
#| eval: false
# === remove numeric rownames introduced by read.csv
if (all(grepl("^[0-9]+$", rownames(F_matrix)))) {
  print("dropping unneeded numeric index")
  # Set the first column as row names
  rownames(F_matrix) <- F_matrix[[1]]
  
  # Remove the first column
  F_matrix <- F_matrix[, -1] 
}
```