One of our customers provided us their teradata workload for analysis.
There are queries like
WHEN t1.c2 IS NOT NULL
AND t1.c2 IS NOT IN ( 'P' , 'PH' ) THEN t1.c4
WHEN t1.c3 IS NOT NULL
AND t1.c3 IS NOT IN ( 'P' , 'PH' ) THEN t1.c5
END AS c6 from t1
We are not able to parse this query since we expected "NOT IN" instead of "IS NOT IN" in our grammar.
I checked the usage manual at page 107 of http://tunweb.teradata.ws/
Can some one please validate if the usage IS NOT IN is correct or not.
This is probably also a leftover of Teradata's pre-SQL query language, Tequel.
Syntatically correct, but proprietary to Teradata and deprecated since at least 20 years :-)