[jira] [Created] (SPARK-28729) Comparison between DecimalType and StringType may lead to wrong results

["ShuMing Li (JIRA)" <ji...@apache.org>]

ShuMing Li created SPARK-28729:
----------------------------------

             Summary: Comparison between  DecimalType and StringType may lead to wrong results
                 Key: SPARK-28729
                 URL: https://issues.apache.org/jira/browse/SPARK-28729
             Project: Spark
          Issue Type: Bug
          Components: SQL
    Affects Versions: 2.3.0
            Reporter: ShuMing Li


 
{code:java}
desc test_table;
a int NULL
b string NULL
dt string NULL
hh string NULL
# Partition Information
# col_name data_type comment
dt string NULL
hh string NULL

select dt from test_table where dt=20190801002382000052000000017638;
20190801002382000052000000017638
20190801002382000052000000017638
20190801002382000052000000016558
{code}
In the sql above, column `dt` is string type. when users forget to add '' in query, Spark returns wrong results.

In `TypeCoercion` class,  DecimalType/StringType is casted as `DoubleType` when DecimalType compares with StringType which maybe not safe with precision lose or truncating.
{code:java}
/**
val findCommonTypeForBinaryComparison: (DataType, DataType) => Option[DataType] = {
....

// There is no proper decimal type we can pick,
// using double type is the best we can do.
// See SPARK-22469 for details.
case (n: DecimalType, s: StringType) => Some(DoubleType)
case (s: StringType, n: DecimalType) => Some(DoubleType)

...
}
{code}
However I cannot find a good solution to avoid this: maybe just throw exception when meets `precision lose` or add a config to avoid this?



--
This message was sent by Atlassian JIRA
(v7.6.14#76016)

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