11
Extras
RMarkdown to R and back again
1
RMarkdown to R and back again
When you have RMarkdown scripts you want to make to R
knitr::purl("01-bakeoff-report.Rmd", documentation = 1)documentation =
An integer specifying the level of documentation to add to the tangled script.
0 means to output pure code, discarding all text chunks);
1 (the default) means to add the chunk headers to the code;
2 means to add all text chunks to code as roxygen comments. roxygen is used in packages
RMarkdown to R and back again
When you have RMarkdown scripts you want to make to R
knitr::purl("01-bakeoff-report.Rmd", documentation = 1)When you have R script you want to make to RMarkdown
❗ Note this will over-write scripts if they have the same name.
knitr::spin("spin-bakeoff-report.R", knit = FALSE, format = "Rmd")documentation =
An integer specifying the level of documentation to add to the tangled script.
0 means to output pure code, discarding all text chunks);
1 (the default) means to add the chunk headers to the code;
2 means to add all text chunks to code as roxygen comments. roxygen is used in packages
Special hash and quote used from royxgen. Shortcuts need to be added manually but can be search and replaced from an R document.
https://github.com/yihui/knitr/blob/master/inst/examples/knitr-spin.R
Also shortcut for comment and uncomment is the same Ctrl + Shift + c
Knitting many documents
2
Knitting many documents
List the Rmd files:
files <- list.files(pattern = "[.]Rmd$")Or list the Rmd files in a folder:
files <- list.files(path = "03-images/", pattern = "[.]Rmd$")Use for loop code to run all the files:
purrr::map( .x = files, .f = ~ rmarkdown::render(glue::glue(.x)))loop in a subfolder
purrr::map( .x = files, .f = ~ rmarkdown::render(glue::glue("03-images/", .x)))Source files
3
Source files
Instead of using import data code in the RMarkdown it can be "sourced" from another script.
source("R/source_file_csv.R")To show this use the 01-bakeoff-report.Rmd and replace the import-data section with this code. The params$series don't need to be removed as this will still work.
Using SQL chunks
4
Using SQL chunks
Using SQL chunks
Show and tell for this SQL code in an RMarkdown
Open file
sql-chunks.Rmd.This file will knit but will just return code - there is no connection.
Add the code
echo = FALSEto the first chunk underDefault connection.
SQL chunks options are supported the same options as R and can also be set globally.
SQL Summary points
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
✔️ Use SQL power R uses RAM so make use of the power of SQL servers
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
✔️ Use SQL power R uses RAM so make use of the power of SQL servers
❗ Each temporary table is an object Unlike in SQL you cannot create a temporary table and then refer to it in later code. You have to use CTEs
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
✔️ Use SQL power R uses RAM so make use of the power of SQL servers
❗ Each temporary table is an object Unlike in SQL you cannot create a temporary table and then refer to it in later code. You have to use CTEs
❗ nvarchar(max) Some tables in SQL use nvarchar(max), varchar(max) or binary which cause your code to fail.
The solution is to do a SELECT statement with the column at the end
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
✔️ Use SQL power R uses RAM so make use of the power of SQL servers
❗ Each temporary table is an object Unlike in SQL you cannot create a temporary table and then refer to it in later code. You have to use CTEs
❗ nvarchar(max) Some tables in SQL use nvarchar(max), varchar(max) or binary which cause your code to fail.
The solution is to do a SELECT statement with the column at the end
❗ Debugging is hard Errors don't come through from SQL so can be hard to debug from RStudio
SQL Summary points
✔️ It's not just R RMarkdown combines text and other code (like SQL)
✔️ Temporary tables - but in R Creating tables of data in R is like creating temporary tables in SQL
✔️ Use SQL power R uses RAM so make use of the power of SQL servers
❗ Each temporary table is an object Unlike in SQL you cannot create a temporary table and then refer to it in later code. You have to use CTEs
❗ nvarchar(max) Some tables in SQL use nvarchar(max), varchar(max) or binary which cause your code to fail.
The solution is to do a SELECT statement with the column at the end
❗ Debugging is hard Errors don't come through from SQL so can be hard to debug from RStudio
❗ Case sensitivity Microsoft SQL is not case sensitive and R is.
And that's the end...
Special thanks to Alison Hill, Emi Tanaka and Desirée De Leon for the original slides and content
More special thanks to NHS-R Community and Hansel Palencia for the RMarkdown workshop run at the 2021 Conference