JARO_WINKLER (Lakehouse v1)

Calculates the Jaro-Winkler distance between two strings. It is commonly used for measuring the similarity between strings, with values ranging from 0.0 (completely dissimilar) to 1.0 (identical strings).

Analyze Syntax

func.jaro_winkler(<string1>, <string2>)

Analyze Examples

func.jaro_winkler('databend', 'Databend')
┌───────────────────────────────────────────┐
│ func.jaro_winkler('databend', 'Databend') │
├───────────────────────────────────────────┤
│                        0.9166666666666666 │
└───────────────────────────────────────────┘

SQL Syntax

JARO_WINKLER(<string1>, <string2>)

SQL Examples

The JARO_WINKLER function returns a FLOAT64 value representing the similarity between the two input strings. The return value follows these rules:

• Similarity Range: The result ranges from 0.0 (completely dissimilar) to 1.0 (identical).

  SELECT JARO_WINKLER('databend', 'Databend') AS similarity;
  
  ┌────────────────────┐
  │     similarity     │
  ├────────────────────┤
  │ 0.9166666666666666 │
  └────────────────────┘
  
  SELECT JARO_WINKLER('databend', 'database') AS similarity;
  
  ┌────────────┐
  │ similarity │
  ├────────────┤
  │        0.9 │
  └────────────┘

• NULL Handling: If either string1 or string2 is NULL, the result is NULL.

  SELECT JARO_WINKLER('databend', NULL) AS similarity;
  
  ┌────────────┐
  │ similarity │
  ├────────────┤
  │ NULL       │
  └────────────┘

• Empty Strings:

  • Comparing two empty strings returns 1.0.

  SELECT JARO_WINKLER('', '') AS similarity;
  
  ┌────────────┐
  │ similarity │
  ├────────────┤
  │          1 │
  └────────────┘

  • Comparing an empty string with a non-empty string returns 0.0.

  SELECT JARO_WINKLER('databend', '') AS similarity;
  
  ┌────────────┐
  │ similarity │
  ├────────────┤
  │          0 │
  └────────────┘