[
https://issues.apache.org/jira/browse/DERBY-2287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12470999
]
Saurabh Vyas commented on DERBY-2287:
-------------------------------------
After some investigation I found that getDigitPrecision() method in
DataTypeUtilities.java returns precision values which is inconsistent with that
returned from metadata.properties. I tried following modification :
public static int getDigitPrecision(DataTypeDescriptor dtd) {
int typeId = dtd.getTypeId().getJDBCTypeId();
switch ( typeId )
{
case Types.FLOAT:
case Types.DOUBLE:
//return TypeId.DOUBLE_PRECISION_IN_DIGITS;
return TypeId.DOUBLE_PRECISION;
case Types.REAL:
//return TypeId.REAL_PRECISION_IN_DIGITS;
return TypeId.REAL_PRECISION;
//Added case for Date & time data type
case Types.DATE:
case Types.TIME:
case Types.TIMESTAMP:
return dtd.getMaximumWidth();
default: return getPrecision(dtd);
}
Then I tried the assert in crossCheckGetColumnRowAndResultSetMetaData() method
of DatabaseMetaDataTest.java which was failing for DATE , REAL and other data
types :
assertEquals("COLUMN_SIZE", rsmdt.getPrecision(col),
rs.getInt("COLUMN_SIZE"));
After above change, DATE & REAL data type are consistent in precision values
and assert but I got assert for JAVACLASSNAME type. This I was not able to root
cause.
Now I have two questions :
- Is my approach for DATE & REAL data type correct
- What is JAVACLASSNAME type & how to handle this.
> JDBC meta data for precision and size is inconsistent and does not match JDBC
> specifications.
> ---------------------------------------------------------------------------------------------
>
> Key: DERBY-2287
> URL: https://issues.apache.org/jira/browse/DERBY-2287
> Project: Derby
> Issue Type: Bug
> Components: JDBC, Network Client
> Affects Versions: 10.0.2.0, 10.0.2.1, 10.1.1.0, 10.1.2.1, 10.1.3.1,
> 10.2.1.6, 10.2.2.0
> Reporter: Daniel John Debrunner
> Priority: Minor
>
> JDBC 4.0 has clarified the definitions of precision in the java doc for
> various methods that return precision or size. The concept of precision and
> size seems to be the same, just having different method or column names in
> various situations.
> Derby does not follow the JDBC 4 specifications consistently, for example -1
> is sometimes used to indicate not applicable, where JBDC 4 says NULL or 0.
> The precision of datetime columns is defined to be non-zero but in some
> situations Derby returns 0.
> jdbcapi.DatabaseMetaDataTest can show some of these issues, the test of
> getColumns() should compare the information in the COLUMN_SIZE column to the
> ResultSetMetaData getPrecision() method for the same column. The comparisions
> are not made currently because the number of mismatches is high. [this code
> is not yet committed].
> Existing application impact as Derby applications may have been relying on
> the old incorrect & inconsistent behaviour.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
