knitr_cawk_engine: knitr engine for the CSV-enhanced Awk language

In hrbrmstr/knitrengines: Collection of Language Engines for Use with the 'knitr' Package

knitr engine for the CSV-enhanced Awk language

Description

The one, true Awk now has support for using CSV columns as fields.

Usage

knitr_cawk_engine(opts)

Arguments

options

knitr options (required parameter)

Details

It will eventually be the default on most distros, until then, to use this knitr engine, you will need to do the following:

• ⁠git clone [e-mail]:onetrueawk/awk.git⁠

• ⁠git checkout csv⁠

• make

• Move a.out to cawk somewhere on your PATH

You can find an example Quarto document that has a cawk section via:

system.file("examples/cawk-test.qmd", package="knitrengines")

Chunk Options

Any chunk with the language of cawk will be processed by this engine. Because there is quite a bit baked into the way Awk works, we need to be able to set up various command line options to get it to work. We do that via knitr chunk options (either inline or in quarto document structured comments).

The following chunk options are supported:

• awk.csv: when set to TRUE, this tells cawk to assume the input files are CSV files. This is not set by default though I'm open to feedback on that life choice.

• awk.var.NAME: CSV-enabled Awk knows nothing about column names. You reference each column by field number. That's not great for "data analysis" work (it isn't great for anything but one-liners, tbh), so one can use the -v CLI option to define variable name mappings to field numbers. So defining awk.var.logdate=1 will let you use ⁠$logdate⁠ in the Awk script vs ⁠$1⁠.

• awk.file.#: Awk processes stdin and files. You can't use data from the previous chunk without shoving it into a file. Awk can also work with multiple input files. You can specify the relative or full path(s) via something like awk.file.1="this.csv". The entire awk.file.# gets ignored. It just needs to be unique. They are passed to the Awk CLI in the order they appear in the options.

Author(s)

Bob Rudis

hrbrmstr/knitrengines documentation built on Sept. 17, 2023, 2:29 a.m.