Create a loop that includes both a code chunk and text

Question

I am trying to figure out how to create a loop that inserts some text into the rmarkdown file, and then produces the graph or table that corresponds to that header. The following is how I picture it working:

for(i in 1:max(month)){
### `r month.name[i]` Air quaility

```{r, echo=FALSE}
plot(airquality[airquality$Month == 5,])
```
}

This ofcourse just prints the for loop as text, if i surround the for loop with r`` I would just get an error.

I want the code to produce an rmd file that looks like this:

May Air Quality

Plot

June Air Quality

Plot

and so on and so forth.

Any ideas? I cannot use latex because I at my work they do not let us download exe files, and I do not know how to use latex anyways. I want to produce a word document.

Answer 1

You can embed the markdown inside the loop using cat().

Note: you will need to set results="asis" for the text to be rendered as markdown. Note well: you will need two spaces in front of the \n new line character to get knitr to properly render the markdown in the presence of a plot out.

# Monthly Air Quality Graphs
```{r pressure,fig.width=6,echo=FALSE,message=FALSE,results="asis"}

attach(airquality)
for(i in unique(Month)) {
  cat("  \n###",  month.name[i], "Air Quaility  \n")
  #print(plot(airquality[airquality$Month == i,]))
  plot(airquality[airquality$Month == i,])
  cat("  \n")
}
```

Answer 2

As mentioned here, you could also make use of the pander package:

# Monthly Air Quality Graphs
```{r pressure2, fig.width=6, echo=FALSE, message=FALSE, results="asis"}
library(pander)
for (i in unique(airquality$Month)) {
   # Inserts Month titles
   pander::pandoc.header(month.name[i], level = 3)
   # Section contents
   plot(airquality[airquality$Month == i,])
   # adding also empty lines, to be sure that this is valid Markdown
   pander::pandoc.p('')
   pander::pandoc.p('')
}
```

Answer 3

What about reusing the chunks inside a loop using <<label>> as described here: https://bookdown.org/yihui/rmarkdown-cookbook/reuse-chunks.html

Label your chunk, set eval=F

 ```{r my_chunk, echo=FALSE, eval=F}
plot(airquality[airquality$Month == 5,])
```

Then loop

for(i in 1:max(month)){
<<my_chunk>>
}

Answer 4

Under some conditions I find it helpful to write a loop that writes chunk code rather than write a chunk that runs a loop. Weird solution but it has worked for me beautifully in the past when a bare bones set of chunks is all I need. For your airquality case it would look like this:

## model chunk ##

# ## May Air Quality
# ```{r May}
# 
# plot(airquality[airquality$Month == 5,])
#
# ```

# all months in airquality
aqmonths <- c("May",
            "June",
            "July",
            "August",
            "September")

for (m in aqmonths) {
  cat(
    paste0(
      "## ", m, " Air Quality",
      "\n\n",
      "```{r ", m, "}",
      "\n\n",
      "plot(airquality[airquality$Month == ", match(m, months), ",])",
      "\n\n",
      "```",
      "\n\n"
    )
  )
}

This will print code for all 5 chunks to the console, and then I can copy and paste into a .Rmd document. It is possible to include any chunk options such as captions or fig arguments in the chunk-writing loop as well. Depending on what else you try to bring in, using functions like match() as in the example is often helpful.

Pros: Preserves ability to use cross-references and set individual captions or options.

Cons: Making changes to all chunks usually requires re-copying the entire output of the chunk-writing loop, which can be tiresome and a bit unwieldy.