csvsort

Sort a CSV file by column 9

$ csvsort [[-c|--columns]] [9] [data.csv]

$ csvsort [[-r|--reverse]] [[-c|--columns]] [name] [data.csv]

$ csvsort [[-c|--columns]] [2,4] [data.csv]

$ csvsort [[-I|--no-inference]] [[-c|--columns]] [columns] [data.csv]

csvsort [options] [FILE]

-c COLUMNS, --columns COLUMNS
    Columns to sort by (repeatable). Use names or 0-based indices; prefix with - for descending.

-r, --reverse
    Reverse (descending) sort order for all columns.

-H, --no-header-row
    Input lacks header row; use indices only.

--basic
    Use basic quoting (minimal escaping).

--delimiter DELIM
    Field delimiter (default: ,)

--doublequote
    Double quotes instead of escape characters.

--escapechar ESCAPE_CHAR
    Escape character (default: ")

--line-terminator LINE_TERMINATOR
    Line ending (default: \n)

--maxfieldsize MAXFIELDSIZE
    Max field length (default: 1048576 bytes).

--quotechar QUOTECHAR
    Quote character (default: ")

--quoting QUOTING
    Quoting style: 0=none, 1=minimal, 2=all, 3=nonnumeric (default: 1).

--skipinitialspace
    Ignore spaces after delimiter.

csvsort is a powerful command-line tool from the csvkit suite for sorting rows in CSV files by one or more columns. It preserves CSV structure, handling quoting, delimiters, and headers intelligently. By default, it treats the first row as a header, excluding it from sorting while using column names for reference.

Specify columns using names (if header present) or zero-based indices. Multiple columns enable multi-level sorts; prefix with - for descending order on that column only (e.g., -c date,-amount sorts date ascending, amount descending). It auto-detects data types like numbers, dates for proper sorting.

Reads from stdin if no file given, ideal for pipelines. Supports custom dialects via options for delimiters, quoting, escapes. Useful for data prep in analysis workflows, before tools like csvstat or csvcut.

Sorting uses Python's stable sorted(), case-sensitive by default. Loads files into memory, so large datasets may require ample RAM.

Loads entire file into memory; unsuitable for huge datasets. Case-sensitive sorting. No built-in stable sort guarantee beyond Python's behavior.

csvsort -c lastname,firstname data.csv
csvsort -c date,-revenue sales.csv
csvcut -c id,name sales.csv | csvsort -c name | csvlook

Names: colname (header required).
Indices: 2 (0-based).
Descending: -2.

Introduced in csvkit 0.9.0 (2011) by Christopher Groskopf; evolved with Python CSV module improvements. Maintained by Kenneth Reitz and community.

csvcut(1), csvlook(1), csvstat(1), csvsort(1), sort(1)