FuzzyMatch

UC Dedicated Cluster14.3+UC Standard ClusterNot SupportedLivyNot Supported

Use the FuzzyMatch gem to identify non-identical duplicates in your data.

Input and Output

DataFrame	Description
in0	Includes the DataFrame on which duplicates will be checked. Note: FuzzyMatch only allows one input.
out	Generates one record per fuzzy match.

Parameters

Configuration

Parameter	Description
Merge/Purge Mode	Records are either compared from a single source (Purge) or across multiple sources (Merge). Merge mode assumes that multiple sources exist in the same DataFrame in0.
Source ID Field	Unique identifier for each source when using Merge mode. This is necessary because the different sources exist in the same DataFrame in0.
Record ID Field	Unique identifier for each record.
Match threshold percentage	If the match score is less than the threshold, the record does not qualify as a match.
Include similarity score	Checkbox to enable for an additional output column that includes the similarity score.

Match Fields

Parameter	Description
Field name	Name of the column that you want to check for duplicates.
Match function	The method that generates the similarity score.

Example

One common use case for the FuzzyMatch gem is to match similarly spelled names. This can be useful for identifying accidentally misspelled names.

1. Create a FuzzyMatch gem and use the customer_id as the Record ID. Then, add a match field for the first_name column.

   

2. Run the gem and see that the output includes the Record IDs of the records with fuzzy matches.

   

3. Join the output with the original dataset to view the matched names.

   

Example code

tip

To see the compiled code of your project, switch to the Code view in the project header.

Python

def fuzzy_match_customers(spark: SparkSession, in0: DataFrame) -> DataFrame:
    from prophecy.utils.transpiler.dataframe_fcns import fuzzyPurgeMode

    return fuzzyPurgeMode(
        in0,
        spark,
        recordId = "customer_id",
        matchFields = {"first_name" : "name"},
        threshold = 0.8,
        includeSimilarityScore = True
    )