Ruby Polars

🔥 Blazingly fast DataFrames for Ruby, powered by Polars

Build Status

Installation

Add this line to your application’s Gemfile:

gem "polars-df"

Getting Started

This library follows the Polars Python API.

Polars.scan_csv("iris.csv")
  .filter(Polars.col("sepal_length") > 5)
  .group_by("species")
  .agg(Polars.all.sum)
  .collect

You can follow Polars tutorials and convert the code to Ruby in many cases. Feel free to open an issue if you run into problems.

Reference

Creating DataFrames

From a CSV

Polars.read_csv("file.csv")

# or lazily with
Polars.scan_csv("file.csv")

From Parquet

Polars.read_parquet("file.parquet")

# or lazily with
Polars.scan_parquet("file.parquet")

From Active Record

Polars.read_database(User.all)
# or
Polars.read_database("SELECT * FROM users")

From JSON

Polars.read_json("file.json")
# or
Polars.read_ndjson("file.ndjson")

# or lazily with
Polars.scan_ndjson("file.ndjson")

From Feather / Arrow IPC

Polars.read_ipc("file.arrow")

# or lazily with
Polars.scan_ipc("file.arrow")

From Avro

Polars.read_avro("file.avro")

From Iceberg (experimental, requires iceberg)

Polars.scan_iceberg(table)

From Delta Lake (experimental, requires deltalake-rb)

Polars.read_delta("./table")

# or lazily with
Polars.scan_delta("./table")

From a hash

Polars::DataFrame.new({
  a: [1, 2, 3],
  b: ["one", "two", "three"]
})

From an array of hashes

Polars::DataFrame.new([
  {a: 1, b: "one"},
  {a: 2, b: "two"},
  {a: 3, b: "three"}
])

From an array of series

Polars::DataFrame.new([
  Polars::Series.new("a", [1, 2, 3]),
  Polars::Series.new("b", ["one", "two", "three"])
])

Attributes

Get number of rows

df.height

Get column names

df.columns

Check if a column exists

df.include?(name)

Selecting Data

Select a column

df["a"]

Select multiple columns

df[["a", "b"]]

Select first rows

df.head

Select last rows

df.tail

Filtering

Filter on a condition

df.filter(Polars.col("a") == 2)
df.filter(Polars.col("a") != 2)
df.filter(Polars.col("a") > 2)
df.filter(Polars.col("a") >= 2)
df.filter(Polars.col("a") < 2)
df.filter(Polars.col("a") <= 2)

And, or, and exclusive or

df.filter((Polars.col("a") > 1) & (Polars.col("b") == "two")) # and
df.filter((Polars.col("a") > 1) | (Polars.col("b") == "two")) # or
df.filter((Polars.col("a") > 1) ^ (Polars.col("b") == "two")) # xor

Operations

Basic operations

df["a"] + 5
df["a"] - 5
df["a"] * 5
df["a"] / 5
df["a"] % 5
df["a"] ** 2
df["a"].sqrt
df["a"].abs

Rounding

df["a"].round(2)
df["a"].ceil
df["a"].floor

Logarithm

df["a"].log # natural log
df["a"].log(10)

Exponentiation

df["a"].exp

Trigonometric functions

df["a"].sin
df["a"].cos
df["a"].tan
df["a"].arcsin
df["a"].arccos
df["a"].arctan

Hyperbolic functions

df["a"].sinh
df["a"].cosh
df["a"].tanh
df["a"].arcsinh
df["a"].arccosh
df["a"].arctanh

Summary statistics

df["a"].sum
df["a"].mean
df["a"].median
df["a"].quantile(0.90)
df["a"].min
df["a"].max
df["a"].std
df["a"].var

Grouping

Group

df.group_by("a").count

Works with all summary statistics

df.group_by("a").max

Multiple groups

df.group_by(["a", "b"]).count

Combining Data Frames

Add rows

df.vstack(other_df)

Add columns

df.hstack(other_df)

Inner join

df.join(other_df, on: "a")

Left join

df.join(other_df, on: "a", how: "left")

Encoding

One-hot encoding

df.to_dummies

Conversion

Array of hashes

df.to_a

Hash of series

df.to_h

CSV

df.to_csv
# or
df.write_csv("file.csv")

Parquet

df.write_parquet("file.parquet")

JSON

df.write_json("file.json")
# or
df.write_ndjson("file.ndjson")

Feather / Arrow IPC

df.write_ipc("file.arrow")

Avro

df.write_avro("file.avro")

Iceberg (experimental)

df.write_iceberg(table, mode: "append")

Delta Lake (experimental)

df.write_delta("./table")

Numo array

df.to_numo

Types

You can specify column types when creating a data frame

Polars::DataFrame.new(data, schema: {"a" => Polars::Int32, "b" => Polars::Float32})

Supported types are:

  • boolean - Boolean
  • decimal - Decimal
  • float - Float16, Float32, Float64
  • integer - Int8, Int16, Int32, Int64, Int128
  • unsigned integer - UInt8, UInt16, UInt32, UInt64, UInt128
  • string - String, Categorical, Enum
  • temporal - Date, Datetime, Duration, Time
  • nested - Array, List, Struct
  • other - Binary, Object, Null, Unknown

Get column types

df.schema

For a specific column

df["a"].dtype

Cast a column

df["a"].cast(Polars::Int32)

Visualization

Add Vega to your application’s Gemfile:

gem "vega"

And use:

df.plot.line("a", "b")

Supports line, pie, column, bar, area, and scatter plots

Group data

df.plot.line("a", "b", color: "c")

Stacked columns or bars

df.plot.column("a", "b", color: "c", stacked: true)

Plot a series

df["a"].plot.hist

Supports hist, kde, and line plots

History

View the changelog

Contributing

Everyone is encouraged to help improve this project. Here are a few ways you can help:

To get started with development:

git clone https://github.com/ankane/ruby-polars.git
cd ruby-polars
bundle install
bundle exec rake compile
bundle exec rake test
bundle exec rake test:docs