SSC-FDM-TD0023

String Similarity might have a different behavior.

Description

This FDM is shown when SnowConvert transforms the Similarity Function from Teradata to Snowflake. It indicates the results might have different behavior.

Example Code

Given the following data as an example

Id	a	b
1
2	Gute nacht	Ich weis nicht
3	Ich weiß nicht	Ich wei? nicht
4	Ich weiß nicht	Ich wei? nicht
5	Ich weiß nicht	Ich weiss nicht
6	Snowflake	Oracle
7	święta	swieta
8	NULL
9	NULL	NULL

Input Code:

Query

IN -> Teradata_01.bteq

SELECT * FROM StringSimilarity (
  ON (
    SELECT id, CAST(a AS VARCHAR(200)) AS a, CAST(b AS VARCHAR(200)) AS b
    FROM table_1
  ) PARTITION BY ANY
  USING
  ComparisonColumnPairs ('jaro_winkler(a,b) AS sim_fn')
  Accumulate ('id')
) AS dt ORDER BY 1;

Result

Id	sim_fn
1	0
2	0.565079365
3	1
4	0.959047619
5	0
6	0.611111111
7	0.7777777777777777
8	0
9	0

Output Code

Query

OUT -> Teradata_01.py

SELECT
* FROM
--** SSC-FDM-TD0023 - STRING SIMILARITY MIGHT HAVE A DIFFERENT BEHAVIOR. **
(
   SELECT
     id,
     JAROWINKLER_UDF(a, b) AS sim_fn
   FROM table_1
 ) dt ORDER BY 1;

Result

ID	SIM_FN
1	0.000000
2	0.560000
3	0.970000
4	0.950000
5	0.000000
6	0.610000
7	0.770000
8	0.000000
9	0.000000

Recommendations

• If you need more support, you can email us at [e-mail]