Unable to use levenshtein function in deterministic rule for Athena Linker

[roshnaramesh]

Hello,
I'm using the Athena linker to perform a link_only comparison between 2 dataframes. I wanted to use the levenshtein() function in the deterministic rule , as mentioned in this example:
deterministic_rules = [ "l.first_name = r.first_name and levenshtein(r.dob, l.dob) <= 1", ] linker.estimate_probability_two_random_records_match(deterministic_rules, recall=0.7)
This is my code:
linker = AthenaLinker(input_table_or_tables=input_tables, boto3_session=boto3_session, output_bucket=output_bucket, output_database=output_database, settings_dict=settings)
deterministic_rules = [ 'levenshtein(r.lastname,l.lastname)<=2' ]
linker.estimate_probability_two_random_records_match(deterministic_rules, recall=0.7)
When I run this, it throws this error
awswrangler.exceptions.QueryFailed: SYNTAX_ERROR: line 29:41: Function levenshtein not registered. You may need to manually clean the data at location '' before retrying. Athena will not delete data in your account.

I'm able to achieve this with the DuckDB linker. Can someone help me with this?

[RossKen]

Hi @roshnaramesh, this is because the levenshtein function is named levenshtein_distance in Athena's SQL engine. https://trino.io/docs/current/functions/string.html

It should work if you change your deterministic rule code