Github user mgaido91 commented on a diff in the pull request:

    https://github.com/apache/spark/pull/23042#discussion_r234155688
  
    --- Diff: 
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/TypeCoercion.scala
 ---
    @@ -138,6 +138,11 @@ object TypeCoercion {
         case (DateType, TimestampType)
           => if (conf.compareDateTimestampInTimestamp) Some(TimestampType) 
else Some(StringType)
     
    +    // to support a popular use case of tables using Decimal(X, 0) for 
long IDs instead of strings
    +    // see SPARK-26070 for more details
    +    case (n: DecimalType, s: StringType) if n.scale == 0 => 
Some(DecimalType(n.precision, n.scale))
    --- End diff --
    
    @cloud-fan I think we have seen many issues on this. I don't think there is 
a standard for them, every RDBMS has different rules. The worst thing about the 
current rules IMHO is that they are not even coherent in Spark (see #19635 for 
instance).
    
    The option I'd prefer is to follow Postgres behavior, ie. no implicit cast 
at all. When there is a type mismatch the user has to choose how to cast the 
things. It is a bit more effort on user side, but it is the safest option IMHO.
    
    What do you think?


---

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

Reply via email to