CODEWITHSUNDEEP
pythonpandasfuzzy-comparison
Ralph Deint
Ralph Deint

Reputation: 380

Match pandas dataframe name columns to another dataframe's columns?

I'm very new to Python. How can I match one text dataframe to another? (kindly please edit this question if I ask this wrongly)

For example given this input data:

 df1 =
          id  Names 
        0 123 Simpson J.
        1 456 Snoop Dogg

 df2 =
            Names 
         0  John Simpson
         1  Snoop Dogg
         2  S. Dogg
         3  Mr Dogg

Is there a way I could find (maybe using findall or match, or any python packages) so that I could produce how many times the names with the id has appeared which almost like this result:

result = 
              id  Names_appeared 
            0 123   1 
            1 456   3

Looking for a brief explanation and some URL to help me understand.

Upvotes: 0

Views: 315

Answers (1)

brennan
brennan

Reputation: 3493

Here's an example using fuzzywuzzy as IanS suggested:

import pandas as pd
from fuzzywuzzy import fuzz


def fuzz_count(shortList, longList, minScore):
    """ Count fuzz ratios greater than or equal to a minimum score. """
    results = []
    for s in shortList:
        scores = [fuzz.ratio(s, l) for l in longList]
        count = sum(x >= minScore for x in scores)
        results.append(count)
    return results


data1 = {'id': pd.Series([123, 456]),
         'Names': pd.Series(['Simpson J.', 'Snoop Dogg'])}
data2 = {'Names': pd.Series(['John Simpson', 'Snoop Dogg', 'S. Dogg', 'Mr Dogg'])}
df1 = pd.DataFrame(data1)
df2 = pd.DataFrame(data2)

result = pd.DataFrame()
result['id'] = df1['id']
counts = fuzz_count(df1['Names'], df2['Names'], minScore=60)  # [1, 2]
result['Names_appeared'] = counts

print(result)  #     id  Names_appeared
               # 0  123               1
               # 1  456               2

Upvotes: 1

Related Questions