In alistaire47/read.so: Read Tables from Stack Overflow Questions into R

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)
Sys.setenv(CLIPR_ALLOW = TRUE)

read.so

Installation

read.so is not on CRAN, but you can install it by running

# install.packages("remotes")
remotes::install_github("alistaire47/read.so")

---

Read data from Stack Overflow questions into R

Sometimes you see a really interesting question on Stack Overflow (or a Github issue, or RStudio Community), but the asker only presents the data as a presentation-style table instead of as runnable R code, because they, unlike you, don't use reprex. Fear no more! read.so will read even heinous tables into data frames and code in a trice.

Read data frame print output back into a data frame

For instance, should you want to return output copied from the R console back into your own session, use read.so for a data.frame, and read_so for a tibble. Pass in a filepath, a raw string of text, a vector of lines, or if the data is on the clipboard, nothing at all, and the functions will grab it for you:

library(read.so)

iris_lines <- capture.output(head(iris))

iris_lines

read.so(iris_lines)

read_so(iris_lines)

clipr::write_clip(head(iris))

read.so()

Further, read_so will attempt to read in the results of printing a tibble:

mtcars_lines <- capture.output(tibble::as_tibble(mtcars))

mtcars_lines

read_so(mtcars_lines)

Read Markdown tables into data frames

When you need to read Markdown tables into R, read.so has you covered with read.md and read_md:

chick_lines <- capture.output(
    knitr::kable(head(ChickWeight), format = "markdown")
)

cat(chick_lines, sep = "\n")

read.md(chick_lines)

read_md(chick_lines)

They can handle a number of formats, including tables with delimiter rows composed of "-", "=", "+", and whitespace.

Read str results back into a data frame

If all you have is the results of calling str on a data frame, read.str will read as many complete rows as possible into a new data frame of the same class as the original:

warp_lines <- capture.output(str(warpbreaks))

warp_lines

read.str(warp_lines)

Read tibble::glimpse results back into a data frame

Similarly, if the data was printed by tibble::glimpse, try read.glimpse or read_glimpse:

states <- data.frame(state.name, state.abb, state.region, state.division, 
                     state.area, center = state.center, state.x77)

states_lines <- capture.output(tibble::glimpse(states))

states_lines

read_glimpse(states_lines)

Generate pretty code to reproduce a data frame

The reading functions make it easy to get a question's data into R, but to construct a reproducible response requires the data in code form. dput will produce runnable code, but the result is not very readable. write.so and its alias write_so take a data frame and restructure the results of dput to make it an ordinary data.frame, tibble, or data.table call, and when a name is detected, reassemble assignment to that variable. The call is printed to the console and by default copied to the clipboard. A language object version of the call is returned, invisibly.

options(read.so.write_clip = FALSE)    # don't write to clipboard by default

write.so(head(swiss))

write_so(head(tibble::as_tibble(swiss)), indent = 2)

Related packages

• reprex: A great way to generate reproducible examples. read.so is [mostly] for when people don't use reprex. You should use reprex.
• datapasta: A broader approach to getting data from the clipboard into R code containing alternatives to write.so. If you like tibble::tribble, you'll like this package.

alistaire47/read.so documentation built on April 23, 2020, 12:50 p.m.