Error there is no package called RevoUtilsMath with R Markdown - r

Question

I am running R 3.4.3 and 3.5.1 (non-Microsoft version) and RStudio version 1.1.456. I am trying to knit some code into RMarkdown. However, I get the following error:
Error in library(p, character.only = TRUE) : there is no package called 'RevoUtilsMath'Calls: <Anonymous> ... suppressPackageStartupMessages -> withCallingHandlers -> library
Execution halted
The package RevoUtilsMath is part of the MKL install with Microsoft R. I cannot install it as a supplemental package with 'regular R'. The script itself runs fine, it just does not work in R Markdown.
The following libraries are loaded:
```{r loadLibraries, echo=FALSE, warning=FALSE}
library(RODBC)
library(dplyr)
library(markovchain)
library(DT)
library(reshape2)
library(knitr)
library(ggplot2)
library(scales)
library(PerformanceAnalytics)
library(plotly)```
The missing package error happens when executing the code below in Markdown. It is called using this code.
```{r histogram1, echo=FALSE, cache=TRUE}```
The histogram1 code is below (very standard ggplot).
g <- ggplot(dataClean, aes(x = IncSnapshotDay, fill = Represent)) +
geom_histogram(bins=70, alpha = .8) +
scale_fill_manual(values = colors) +
scale_x_continuous(labels = comma, limits = c(0,40000)) +
facet_wrap(~SnapshotDay) +
ylim(0,4000) +
theme_bryan()
g
I thought that maybe one of these packages has a dependency, so I ran the following to find out.
library(tools)
> dependsOnPkgs('RevoUtilsMath')
It returns character(0) which indicates that none of the packages depend on it. I did a test of the function on ggplot2, and it works
dependsOnPkgs('ggplot2')
[1] "dendextend" "GGally" "ggthemes" "plotly" "viridis" "caret" "crosstalk"
[8] "DT"
So why does R Markdown/knitr generate this error since the code itself runs fine outside of Markdown, and how do I fix this?

Answer 1

I think you are experiencing a problem similar to one I just had though, without more detail, it is hard for me to know.
In my case, the problem was caused by knitr caching the list of packages used by a previous author (using Microsoft R). The immediate solution was to simply clear the knitr cache (via the "Knit" drop down menu in RStudio) before attempting to knit the code.
I still don't really understand why this is happening or how to avoid it in future situations, but this at least provides a way to create the document even in the face of this behavior.