a-pandas-ex-fuzzymerge 0.10

Merges two DataFrames using fuzzy matching on specified columns

Merges two DataFrames using fuzzy matching on specified columns

Tested against Windows / Python 3.11 / Anaconda

pip install a-pandas-ex-fuzzymerge

This function performs a fuzzy matching between two DataFrames `df1` and `df2`
based on the columns specified in `right_on` and `left_on`. Fuzzy matching allows
you to find similar values between these columns, making it useful for matching
data with small variations, such as typos or abbreviations.

Parameters:
df1 (DataFrame): The first DataFrame to be merged.
df2 (DataFrame): The second DataFrame to be merged.
right_on (str): The column name in `df2` to be used for matching.
left_on (str): The column name in `df1` to be used for matching.
usedtype (numpy.dtype, optional): The data type to use for the distance matrix.
	Defaults to `np.uint8`.
scorer (function, optional): The scoring function to use for fuzzy matching.
	Defaults to `fuzz.WRatio`.
concat_value (bool, optional): Whether to add a 'concat_value' column in the result DataFrame,
	containing the similarity scores. Defaults to `True`.
**kwargs: Additional keyword arguments to pass to the `pandas.merge` function.

Returns:
DataFrame: A merged DataFrame with rows that matched based on the specified fuzzy criteria.

Example:
	from a_pandas_ex_fuzzymerge import pd_add_fuzzymerge
	import pandas as pd
	import numpy as np
	from rapidfuzz import fuzz
	pd_add_fuzzymerge()
	df1 = pd.read_csv(
		"https://raw.githubusercontent.com/pandas-dev/pandas/main/doc/data/titanic.csv"
	)
	df2 = df1.copy()
	df2 = pd.concat([df2 for x in range(3)], ignore_index=True)
	df2.Name = (df2.Name + np.random.uniform(1, 2000, len(df2)).astype("U"))
	df1 = pd.concat([df1 for x in range(3)], ignore_index=True)
	df1.Name = (df1.Name + np.random.uniform(1, 2000, len(df1)).astype("U"))

	df3 = df1.d_fuzzy_merge(df2, right_on='Name', left_on='Name', usedtype=np.uint8, scorer=fuzz.partial_ratio,
							concat_value=True)
	print(df3)