phrs — the tabular data shell tool

phrs is a Rust command-line tool for manipulating CSV data in Unix pipelines. It is inspired by ph and uses Polars for tabular operations.

Install

cargo install phrs

Getting started

Suppose a.csv contains:

x,y
3,8
4,9
5,10
6,11
7,12
8,13

Commands usually read CSV from standard input and write CSV to standard output, so they can be chained:

cat a.csv | phrs columns y x | phrs head 3

y,x
8,3
9,4
10,5

To display a human-readable table, use show:

cat a.csv | phrs show

      x    y
--  ---  ---
 0    3    8
 1    4    9
 2    5   10
 3    6   11
 4    7   12
 5    8   13

show produces terminal output rather than CSV; it is normally the last command in a pipeline.

Commands

Inspect and select columns

cat a.csv | phrs columns          # List column names
cat a.csv | phrs columns y x      # Select and reorder columns
cat a.csv | phrs shape            # Output row and column counts as CSV

Select and order rows

cat a.csv | phrs head             # First 10 rows (default)
cat a.csv | phrs head 3
cat a.csv | phrs tail             # Last 10 rows (default)
cat a.csv | phrs tail 3
cat a.csv | phrs sort x           # Sort by column x

Rename and clean columns

cat a.csv | phrs rename x number  # Rename column x to number
cat a.csv | phrs slugify          # Normalize column names
cat a.csv | phrs strip            # Trim whitespace in string columns
cat a.csv | phrs strip x          # Trim a specified column

Filter rows with query

query supports simple column-to-value comparisons (==, !=, <, <=, >, >=):

cat a.csv | phrs query 'x>5'
cat a.csv | phrs query 'y <= 10'

Compute columns with eval

eval supports arithmetic expressions, assignments, numeric constants, parentheses, and exponentiation:

cat a.csv | phrs eval 'z = x + y'
cat a.csv | phrs eval 'z = x**2 + y'
cat a.csv | phrs eval 'z = (x + y) / 2'
cat a.csv | phrs eval 'x**2'       # Output only the computed column

Display a table

cat a.csv | phrs show
cat a.csv | phrs show --no-index  # Omit the generated row index

Concatenate CSV files

Unlike most commands, cat takes filenames. With no filenames, it reads standard input.

phrs cat a.csv b.csv
phrs cat a.csv b.csv --axis=index
phrs cat a.csv b.csv --axis=columns
cat a.csv | phrs cat

Merge CSV files

merge reads two named CSV files and joins their rows. The default join type is inner; supported types are inner, left, right, and outer.

phrs merge left.csv right.csv
phrs merge left.csv right.csv --how=outer
phrs merge left.csv right.csv --on=key
phrs merge left.csv right.csv --left=left_key --right=right_key

When no join key is provided, common column names are used as join keys.

Pipelines

cat a.csv \
  | phrs query 'x>4' \
  | phrs eval 'z = x * y' \
  | phrs columns x z \
  | phrs sort x \
  | phrs show

Development

cargo test
cargo build --release

The optimized executable is written to target/release/phrs.

phrs aims to reproduce the behavior of the Python ph tool for its implemented commands, but does not yet support every ph operation. Differences are tracked using differential tests against the Python implementation.

License

MIT.