[GitHub] [spark] cloud-fan commented on issue #27965: [SPARK-31205][SQL] support string literal as the second argument of date_add/date_sub functions

[GitBox <gi...@apache.org>]

cloud-fan commented on issue #27965: [SPARK-31205][SQL] support string literal as the second argument of date_add/date_sub functions
URL: https://github.com/apache/spark/pull/27965#issuecomment-601657783
 
 
   @yaooqinn the reason why almost all the mainstream databases accept string literals as SQL function arguments is because it's convenient and common. Some BI tools even generate SQL queries with string literals in function arguments regardless of the required type.
   
   Spark is mostly compatible with this use case because our type coercion can implicitly cast string to other types. This is bad as it can lead to runtime null values for invalid string values, and we are trying to improve it under the ANSI mode.
   
   #26412 breaks this use case. I'm OK to add back string type support completely, but I do agree the previous behavior was very confusing. This PR proposes a temporary workaround to support valid string literals only, and I believe we will have a holistic solution of ANSI type coercion in the next release, which handles string literals and invalid string values.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org