Skip to content

"nt" tables utility design

ilyash-b edited this page Jul 26, 2018 · 1 revision

Background

The most common data format when using command line is tables. Unfortunately, almost no CLI tools process tables.

Existing tools in the vicinity

awk

https://www.gnu.org/software/gawk/manual/gawk.html

Note that awk processes records, not tables. Using $1 and so on to reference a column is not ideal.

parallel

https://www.gnu.org/software/parallel/

tabulate

https://bitbucket.org/astanin/python-tabulate

Problem

There are many situations in which tabular data manipulation is required.

Requirements

  • Input formats support (TODO: add more)
    • JSON lines
  • "nt" must be able to process the data as a stream (as opposed to loading all the data into memory first)

Design

TODO