pychisel 0.2

Automated tools for quick preparation of Pandas DataFrames.

PyChisel

PyChisel is a Python package that provides automated tools for quick preparation of Pandas DataFrames. It simplifies the process of data normalization by splitting columns into reference tables, which is a common technique in data preparation and database design.

Installation

You can install PyChisel using pip:

pip install pychisel

Requirements

• Python 3.11 or higher
• pandas 2.0.0 or higher

Features

• Column Splitting: Split columns in a DataFrame into reference tables, replacing original values with IDs
• Customizable Configuration: Control how reference tables and ID columns are named
• Inplace Operations: Choose whether to modify the original DataFrame or create a copy

Usage

Basic Usage

import pandas as pd
from pychisel.core import Splitter

# Create a sample DataFrame
df = pd.DataFrame({
    'city': ['New York', 'Los Angeles', 'Chicago', 'New York', 'Los Angeles'],
    'department': ['Sales', 'Marketing', 'Sales', 'IT', 'Sales'],
    'employee_count': [10, 15, 8, 12, 20]
})

# Initialize the Splitter
splitter = Splitter()

# Split the 'city' and 'department' columns
splitted_df, ref_tables = splitter.split(df, columns=['city', 'department'])

# The original DataFrame now has ID columns instead of the original values
print(splitted_df)

# Reference tables contain the mapping between IDs and original values
print(ref_tables['ref_city'])
print(ref_tables['ref_department'])

Using the Convenience Function

import pandas as pd
from pychisel.core import split

# Create a sample DataFrame
df = pd.DataFrame({
    'city': ['New York', 'Los Angeles', 'Chicago', 'New York', 'Los Angeles'],
    'department': ['Sales', 'Marketing', 'Sales', 'IT', 'Sales'],
    'employee_count': [10, 15, 8, 12, 20]
})

# Split the 'city' and 'department' columns using the convenience function
splitted_df, ref_tables = split(df, columns=['city', 'department'])

Advanced Configuration

import pandas as pd
from pychisel.core import Splitter

# Create a sample DataFrame
df = pd.DataFrame({
    'city': ['New York', 'Los Angeles', 'Chicago', 'New York', 'Los Angeles'],
    'department': ['Sales', 'Marketing', 'Sales', 'IT', 'Sales'],
    'employee_count': [10, 15, 8, 12, 20]
})

# Initialize the Splitter with custom configuration
splitter = Splitter(
    drop_original=False,  # Keep the original columns
    ref_prefix='dim_',    # Use 'dim_' as prefix for reference tables
    id_suffix='_key'      # Use '_key' as suffix for ID columns
)

# Split the 'city' and 'department' columns
splitted_df, ref_tables = splitter.split(df, columns=['city', 'department'])

# The DataFrame now has both original columns and ID columns
print(splitted_df)

# Reference tables are named with the custom prefix
print(ref_tables['dim_city'])
print(ref_tables['dim_department'])

License

This project is licensed under the MIT License.

Author

Anderson Alves Monteiro - GitHub