Thanks Greg, for getting on that last one so fast. I'm on my way with
this Oracle DATE/TIMESTAMP issue. I still cannot find the Village patch
that is mentioned in mailing list archives, so I am rolling my own. I'm
still not sure if patching Village is the way to go, or if I should just
use the -Doracle.jdbc.V8Compatible=true JVM paramter mentioned in the
Wiki.
If Village was to be patched, is this totally crazy?
Index: src/java/com/workingdogs/village/Column.java
===================================================================
--- src/java/com/workingdogs/village/Column.java (revision 21344)
+++ src/java/com/workingdogs/village/Column.java (working copy)
@@ -139,6 +139,17 @@
this.nullAllowed = rsmd.isNullable(columnNumber) == 1;
this.autoIncrement = rsmd.isAutoIncrement(columnNumber);
+ // ORACLE DATE BUGFIX / HACK
+ // Oracle 9.2+ returns DATE SQL type for DATE columns, even though it
+ // can store hour/minute/second info. Bend Oracle DATE columns to
+ // make Village believe they are timestamp fields
+ if (rsmd.getClass().getName().equals(
+ "oracle.jdbc.driver.OracleResultSetMetaData") &&
+ this.columnType == Types.DATE) {
+ this.columnTypeName = "TIMESTAMP";
+ this.columnType = Types.TIMESTAMP;
+ }
+
// The JDBC spec is VERY unclear about what this means and as
// such, it should be ignored. Derby returns true all the time.
// Sybase and Informix say it's unsupported (and false).
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
