I have to make a correction to the last statement in my original post:
The content values of DISCOUNT_CODE field are /*not*/ of numeric value,
but actual characters, such as 'H' or 'M'. Sorry for causing confusion.
On 10/24/2013 03:17 PM, Mark Struberg wrote:
My finding so far is that the DerbyDictionary doesn't overwrite
storeCharsAsNumbers, so it's still defaulted to true.
This code is exactly which trashes in the sample
Caused by: java.sql.SQLDataException: Invalid character string format for type
int.
at
org.apache.derby.client.am.SQLExceptionFactory40.getSQLException(Unknown Source)
at org.apache.derby.client.am.SqlException.getSQLException(Unknown
Source)
at org.apache.derby.client.am.ResultSet.getInt(Unknown Source)
at
org.apache.commons.dbcp.DelegatingResultSet.getInt(DelegatingResultSet.java:225)
at
org.apache.commons.dbcp.DelegatingResultSet.getInt(DelegatingResultSet.java:225)
at
org.apache.openjpa.lib.jdbc.DelegatingResultSet.getInt(DelegatingResultSet.java:137)
at
org.apache.openjpa.jdbc.sql.DBDictionary.getInt(DBDictionary.java:841)
at
org.apache.openjpa.jdbc.sql.DBDictionary.getChar(DBDictionary.java:740)
at
org.apache.openjpa.jdbc.sql.ResultSetResult.getCharInternal(ResultSetResult.java:300)
at
org.apache.openjpa.jdbc.sql.ResultSetResult.getObjectInternal(ResultSetResult.java:377)
The stacktrace is with todays 2.3.x branch snapshot.
DBDictionary#getChar():
public char getChar(ResultSet rs, int column)
throws SQLException {
if (storeCharsAsNumbers)
return (char) getInt(rs, column); <-- line 740 boom
...
The question for me is how Derby usually stores char fields internally.
Is our DerbyDictionary just wrong and should it set storeCharsAsNumbers to
false?
There is of course the option to use @ColumnDefinition in the entity. This
should bypass this flag (I hope, didn't test it yet).
LieGrue,
strub
----- Original Message -----
From: Kay Wrobel <kay.wro...@gmx.net>
To: dev@openjpa.apache.org
Cc:
Sent: Thursday, 24 October 2013, 22:08
Subject: PersistenceException: Invalid character string format for type int
Dear OpenJPA Dev Team,
I am getting this exception
<https://raw.github.com/kwrobel/OpenJPA-IntExceptionTest/master/OpenJPA-IntExceptionTest/DiscountCode-Exception.txt>
while reading a simple entity DiscountCode from a Derby database called
SAMPLE (ships with Glassfish/NetBeans) called DISCOUNT_CODE that has
only two fields:
1. DISCOUNT_CODE CHAR(1) NOT NULL PRIMARY KEY
2. RATE DECIMAL(4,2) NULL
DISCOUNT_CODE maps to type "Character" and RATE maps to
"BigDecimal". I
set up a Github <https://github.com/kwrobel/OpenJPA-IntExceptionTest>
with a small test project. It's really small and concise. All you need
to run this is a running Derby Network server with the SAMPLE database
(included on the github), the latest OpenJPA library (I tested so far:
2.2.0 shipping with TomEE 1.5.2, 2.2.2 from the main project page, and
2.3.0-SNAPSHOT), the derby client driver (derby.jar, derbyclient.jar,
derbynet.jar from the main Apache Derby distribution) and that's it.
The curious part is that the field DiscountCode, though a char field,
contains numeric, integer values. Mark Struberg seems to have a hunch
already of what is going on and is able to add some more insight.
Any input on this is greatly appreciated.
Regards,
Kay Wrobel
