好吧，一年半过去了，我仍然不时需要这个，但从那时起我学到了足够多的知识，知道如何正确地做到这一点。您需要编写一个knitr“输出挂钩”，并修改输出，以便 html 可以“逃脱”。
以下实现：
- 添加了
knitr 输出挂钩。
- 定义了一个语法来指定所需的标签和内容
- 例如获取
<h1>some_text</h1> 使用htmlesc<<(h1,some_text)>>
- 想出一个正则表达式来提取
h1 和some_text 并将其重新格式化为正确标记的html，同时删除knitr 插入的##。
- 添加了更多测试用例以确保它执行了一些额外的操作（例如
h4、p、正确放置图表和表格等）
- 添加了另一个正则表达式来删除双转义行，这会添加一些不需要的空格面板结构。
代码如下：
---
title: "Output Hook for HTML Escape"
author: "Someone"
date: "2017 M04 25"
output:
html_document:
keep_md: true
---
```{r setup, include=T,echo=TRUE}
knitr::opts_chunk$set(echo = TRUE)
hook_output <- knitr::knit_hooks$get("output")
knitr::knit_hooks$set(output=function(x,options){
xn <- hook_output(x,options)
# interestingly xn is a big character string.
# I had expected a list or vector, but the length of x is 1 and the class is character.
# The following regexp extracts the parameters from a statement of the form
# htmlesc<<(pat1,pat2)>> and converts it to <pat1>pat2</pat1>
xn <- gsub("## htmlesc<<([^,]*),{1}([^>>]*)>>","\n```\n<\\1>\\2</\\1>\n```\n",xn)
# now remove double escaped lines that occur when we do these right after each other
gsub(">\n```\n\n\n```\n<",">\n<",xn)
}
)
```
## An analysis loop in a single R chunk with R Markdown
In the following, we do a loop and generate 3 sets of data:
(@) We have some explanitory text
(@) then we do a bar plot with ggplot
(@) then we print out a table
(@) then we do a base plot - just for fun
```{r, echo=T, fig.height=3,fig.width=5}
library(knitr)
library(tidyr)
library(ggplot2)
set.seed(123)
for (i in 1:3){
mdf <- data.frame(matrix(runif(25),5,5))
cat(sprintf("htmlesc<<h1,Title %d>>\n",i))
cat(sprintf("htmlesc<<h4,Smaller Title - also for %d>>\n",i))
cat(sprintf("htmlesc<<p,and some text talking about this %d example>>\n",i))
print(sapply(mdf,mean))
gdf <- gather(mdf,series,val)
gp <- ggplot(gdf)+geom_bar(aes(series,val,fill=series,color=I("black")),stat="identity")
print(gp)
print(mdf)
plot(mdf)
}
```
这是输出（缩小了一点，因为您不需要详细信息）。
顺便说一句，唯一真正的文档是易辉的优秀编织书，搜索很容易找到。