ColumnPairTrends score depends on the data index

[R-Palazzo]

Environment Details

Please indicate the following details about the environment in which you found the bug:

• SDMetrics version: 0.14.1

Error Description

By design, the metrics and property score should be independent of the indexes of the real and synthetic data. However, this is currently not the case for the ColumnPairTrends property as shown below. The issue comes from the discretization step when numerical and datetime columns are converted to categorical, the indexes are not preserved.

Steps to reproduce

The code below should output a metric score of 1.0 since real and synthetic data are the same (the only have different indexes).

import pandas as pd
from sdmetrics.reports.single_table._properties import ColumnPairTrends

real_data = pd.DataFrame({
    'A': [1, 2, 3],
    'B': ['a', 'b', 'a']
}, index=[0, 1, 2])

synthetic_data = pd.DataFrame({
    'A': [1, 2, 3],
    'B': ['a', 'b', 'a']
}, index=[0, 4, 2])

metadata = {
    'columns': {
        'A': {'sdtype': 'numerical'},
        'B': {'sdtype': 'categorical'}
    }
}

property = ColumnPairTrends()
property._generate_details(real_data, synthetic_data, metadata)

The current output is: