miller

TLDR

Convert CSV to JSON

$ mlr --csv --ojson cat [file.csv]

Filter records

$ mlr --csv filter '$[field] > [10]' [file.csv]

Select columns

$ mlr --csv cut -f [name,age] [file.csv]

Sort by field

$ mlr --csv sort -f [age] [file.csv]

Aggregate statistics

$ mlr --csv stats1 -a sum,mean -f [value] -g [category] [file.csv]

Transform field

$ mlr --csv put '$[total] = $[price] * $[quantity]' [file.csv]

SYNOPSIS

mlr [options] verb [verb-options] files...

Miller (mlr) is like awk, sed, cut, join, and sort for name-indexed data such as CSV, TSV, and JSON. It's designed for processing structured text data from the command line.
Miller handles heterogeneous data formats with consistent syntax.

PARAMETERS

--csv

CSV input format.

--json

JSON input format.

--ojson

JSON output format.

--icsv --ojson

CSV in, JSON out.

cat

Pass through records.

filter

Select matching records.

cut

Select fields.

sort

Sort records.

put

Add/modify fields.

stats1

Aggregate statistics.

FORMAT CONVERSION

$ mlr --icsv --ojson cat data.csv
mlr --ijson --ocsv cat data.json
mlr --csv --opprint cat data.csv

CAVEATS

DSL syntax differs from awk. Field names are case-sensitive. Large files processed streaming.

HISTORY

Miller was created by John Kerl starting in 2015 to provide a unified tool for structured text processing.