Fix unsigned jdbc types

[panbingkun]

1.The pr aims to fix unsigned jdbc types, include:

• UnsignedIntColumn from Types.BIGINT to Types.INTEGER
• UnsignedSmallIntColumn from Types.INTEGER to Types.SMALLINT
• UnsignedTinyIntColumn from Types.SMALLINT to Types.TINYINT

2.why?
When we recently migrated mariadb-connector-j from 2.x to 3.x in Spark and used mariadb-connector-j to access MySQL data, We found that 'java. sql.Types' returned by some unsigned jdbc types of '2.x ' and ' 3.x ' are inconsistent, as follows:

Column Type Define	2.x	3.x
TINYINT UNSIGNED	Types.TINYINT	Types.SMALLINT
SMALLINT UNSIGNE	Types.SMALLINT	Types.INTEGER
INT UNSIGNED	Types.INTEGER	Types.BIGINT

val dataType = resultSetMetaData.getColumnType(i)

3.The data type is generally determined by two attributes: ColumnType and isSigned

[panbingkun]

cc @rusher @vaintroub @massimosiani
In addition, this has been the case since the beginning of the 3.x version (eg: 3.0.3).
Please don't mind. I want to ask, is there any special consideration or is it a bug?

The pr of migration mariadb-connector-j corresponding to Spark is: apache/spark#46655

[panbingkun]

also cc @yaooqinn @dongjoon-hyun