pandas-term 0.1.0

CLI bringing pandas operations to the command line

pandas-term

pandas-term is a CLI bringing pandas operations to the command line.

Note: Still in early experimental development and may change

Installation

pipx install pandas-term

or

uv tool install pandas-term

Usage

All commands accept an input file path (or - for stdin) and support -o/--output for file output (default: stdout).

Command Reference

Command	Pandas Equivalent	Description
pd select	df[columns]	Select columns
pd drop	df.drop()	Drop columns
pd rename	df.rename()	Rename columns
pd sort	df.sort_values()	Sort by columns
pd dedup	df.drop_duplicates()	Remove duplicate rows
pd merge	pd.merge()	Merge two dataframes
pd concat	pd.concat()	Concatenate dataframes
pd batch	df.iloc[]	Split into batches
pd query	df.query()	Filter with query expr
pd head	df.head()	First n rows
pd tail	df.tail()	Last n rows
pd dropna	df.dropna()	Drop rows with nulls
pd describe	df.describe()	Descriptive statistics
pd unique	df[col].unique()	Unique values in column
pd shape	df.shape	Dimensions (rows, columns)
pd columns	df.columns	Column names
pd dtypes	df.dtypes	Column data types
pd value-counts	df.value_counts()	Count unique values
pd groupby	df.groupby().agg()	Group by and aggregate

Transform

# Select columns
pd select name,age data.csv

# Drop, sort & rename
pd drop unwanted_column data.csv
pd sort age data.csv --descending
pd rename "name:full_name,age:years" data.csv

# Remove duplicates
pd dedup data.csv
pd dedup --subset name,email data.csv

# Merge two dataframes
pd merge left.csv right.csv --on user_id --how inner

# Concatenate files (supports glob patterns)
pd concat file1.csv file2.csv
pd concat "data_*.csv"

# Split into batches
pd batch data.csv --sizes 100 -o "batch_{}.csv"

Filter

# Query expression
pd query "age > 30 and city == 'NYC'" data.csv

# First/last N rows
pd head --n 100 data.csv
pd tail --n 50 data.csv

# Drop rows with nulls
pd dropna data.csv
pd dropna --subset "name,age" data.csv

Stats

pd describe data.csv
pd unique country data.csv
pd shape data.csv
pd columns data.csv
pd dtypes data.csv

Aggregate

# Count unique values
pd value-counts city data.csv
pd value-counts city,department data.csv --normalize

# Group by and aggregate
pd groupby department data.csv --col salary --agg sum
pd groupby "city,department" data.csv --col salary,age --agg mean

Piping

All commands support piping through stdin/stdout. When piping, you can omit the input file argument (it defaults to stdin):

cat data.csv | pd query "age > 30" | pd select name,age

pd sort age data.csv --descending | pd head --n 10 | pd select name,age

Output Formats

Use -f/--format for stdout format (default: csv):

pd head --n 10 data.csv -f json
pd head --n 10 data.csv -f tsv
pd head --n 10 data.csv -f md

--json/-j is shorthand for -f json.

File output format is determined by extension:

pd select name,age data.csv -o output.xlsx
pd query "age > 30" data.json -o filtered.parquet

Supported: csv, tsv, json, xlsx, parquet, md

For other extensions, use redirection: pd select name data.csv -f csv > output.txt

Dev setup

Requires uv

Install dependencies:

uv sync

Command	Description
make check	Format, lint, and test
make format	Format code
make lint	Linting only
make test	Run tests
make snapshots	Regenerate test snapshots
make coverage	Tests with coverage
make bump	Bump version number
make demo	Regnerate demo gif from tape file