Implementation:Evidentlyai Evidently Legacy Exact Match Feature

Knowledge Sources	Evidentlyai_Evidently
Domains	ML Monitoring, Text Comparison
Last Updated	2026-02-14 12:00 GMT

Overview

Computes an exact equality comparison between two DataFrame columns, producing a categorical boolean feature indicating whether the values in each row match.

Description

The ExactMatchFeature class extends GeneratedFeature to perform a straightforward element-wise equality check between two columns specified in the columns list. The generate_feature() method compares data[columns[0]] == data[columns[1]] and returns a DataFrame with a single boolean column.

The feature name is constructed by joining the two column names with a pipe character (e.g., "prediction|reference"), and the default display name follows the pattern "Exact Match for {col1} {col2}." The feature type is Categorical since the result is a boolean True/False value.

This is one of the simplest comparison features in the Evidently feature library, useful as a baseline text or value comparison metric before applying more sophisticated similarity measures.

Usage

Use this feature when you need a simple binary check of whether two columns contain identical values row-by-row. Common use cases include comparing model predictions against ground truth labels, checking if generated text exactly matches a reference, or validating data consistency between duplicate columns.

Code Reference

Source Location

• Repository: Evidentlyai_Evidently
• File: src/evidently/legacy/features/exact_match_feature.py

Signature

class ExactMatchFeature(GeneratedFeature):
    class Config:
        type_alias = "evidently:feature:ExactMatchFeature"

    __feature_type__: ClassVar = ColumnType.Categorical
    columns: List[str]

    def generate_feature(self, data: pd.DataFrame, data_definition: DataDefinition) -> pd.DataFrame: ...
    def _feature_name(self): ...
    def _as_column(self) -> ColumnName: ...

Import

from evidently.legacy.features.exact_match_feature import ExactMatchFeature

I/O Contract

Inputs

Name	Type	Required	Description
columns	List[str]	Yes	A list of exactly two column names to compare for exact equality.
data	pd.DataFrame	Yes (at generation time)	The input DataFrame containing both columns.
data_definition	DataDefinition	Yes (at generation time)	The data definition describing the dataset schema.

Outputs

Name	Type	Description
return	pd.DataFrame	A single-column DataFrame with boolean values (True where values match, False otherwise). Column name is the pipe-joined input column names.

Usage Examples

from evidently.legacy.features.exact_match_feature import ExactMatchFeature

# Compare prediction against ground truth
exact_match = ExactMatchFeature(columns=["prediction", "ground_truth"])

# Generate the feature
# result_df = exact_match.generate_feature(data=my_df, data_definition=data_def)
# result_df will contain a column "prediction|ground_truth" with True/False values

# Get the column reference for use in metrics
column_ref = exact_match.as_column()

Related Pages

• Environment:Evidentlyai_Evidently_Python_Core_Environment
• Evidentlyai_Evidently_Legacy_Generated_Features - Base class GeneratedFeature that ExactMatchFeature extends.
• Evidentlyai_Evidently_Legacy_BERTScore_Feature - A more sophisticated semantic similarity feature for text comparison.