How to generate multiple files from r script function with knitr

Question

I would like to generate a number of reports from an r script. I would like to avoid the duplication needed when compiling from rmd files and I'd like to generate the reports from the r script itself if possible.

The script has a function which should generate each report from a list of dataframes where each report represents the processing of one dataframe. I just can't understand how to generate each report separately. At the moment the report shows all of the dataframes in one file

The original script is long so I have provided a minimal version below.

    structure(list(a = structure(c(1L, 3L, 2L), .Label = c("boo", 
    "saa", "yaew"), class = "factor"), b = structure(c(2L, 3L, 
    1L), .Label = c("mfds", "shu", "ren"), class = "factor"), c = structure(c(2L, 
    1L, 3L), .Label = c("22", "23", "5345"), class = "factor")), .Names = c("a", 
    "b", "c"), row.names = c(NA, -3L), class = "data.frame")

    ReportOp<-function(n) { 
  this_is_a_name <- n; 
  this_is_my_data <- ldf[[n]] 

  #' ---
  #' author: Me
  #' date:
  #' ---

  #+results='asis', echo=FALSE
  knitr::kable(this_is_my_data, digits = 2)
  #'
} 

At the moment I am just generating the report with everything in it using thecompile notebook button in R studio.

I tried using knitr::spin on the above script as follows from a separate file:

library(knitr)
o=spin('/Users/sebastianzeki/Desktop/UntitledTBB.R')
knit2html(o,output="/Users/sebastianzeki/Desktop/out.html")

but again I only get one report outputted rather than three.

Answer 1

You would make a single rmarkdown parameterized report. Define the parameters in the YAML. You render the report by passing the parameters to rmarkdown::render() in a list.

You'd run this over multiple sets of parameters by various means in r, loops, the apply family (mapply in this case), dplyr::do, or the purrr:map() functions. In this example, I use purrr::pmap().

Report.Rmd

---
title: "`r sprintf('mtcars %s vs. %s', params$variable1, params$variable2)`"
output: html_document
params:
  variable1: "mpg"
  variable2: "cyl"
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
library(ggplot2)
```
#Plot
Crazy report text!

```{r cars}
ggplot(mtcars, aes_(x = as.name(params$variable1), y = as.name(params$variable2))) +
  geom_point()
```

Render Function

render_report <- function(var1, var2) {

  template <- "path_to/Report.Rmd"

  out_file <- sprintf("report %s vs. %s", var2, var1)

  parameters <- list(variable1 = var1,
                     variable2 = var2)

    rmarkdown::render(template,
                      output_file = out_file,
                      params = parameters)
  invisible(TRUE)
}

Running over multiple parameters.

library(purrr)

params_list <- list(list("mpg","mpg","mpg"),
                    list("drat","wt","qsec"))

pmap(params_list, render_report)

Output

Three html files, each named by the parameters, each with the scatterplot according to the parameters.

1. drat vs mpg

2. qsec vs mpg

3. wt vs mpg