Mailing-List: contact derby-dev-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Development" <derby-dev@db.apache.org>
Received-SPF: pass (hermes.apache.org: local policy)
Message-ID: <41E71EE5.8020809@golux.com>
Date: Thu, 13 Jan 2005 17:22:45 -0800
From: Army <army@golux.com>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US;
 rv:1.7.1) Gecko/20040707
MIME-Version: 1.0
To: Derby Development <derby-dev@db.apache.org>
Subject: Date formatting with Network Server
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Question regarding date formatting in the Network Server:

	System.out.println(java.sql.Date.valueOf("0001-01-01"));

With a Sun JVM, the above line will print "0001-01-01".  With an IBM JVM, it will print "1-01-01".  The difference is 
apparently in the implementation of the "toString()" method for the two JVMs.

Currently, when a query against Network Server returns a date column, the string value for that column is returned using 
the following line (in DRDAConnThread.java):

	writer.writeString(((java.sql.Date) val).toString());

This works fine for Sun JVM, because the toString() method returns "0001-01-01".  However, for IBM JVM, the string 
"1-01-01" is returned, and that causes the JDBC client to fail, presumably because the client sees it as an invalid date 
string.

This is reproduced easily enough; start the server with an IBM JVM, connect to it using ij, insert the value 
"0001-01-01" into a table, then select from the table.  You'll get an error.

So my question is three-fold:

1) The Java definition for java.sql.Date.toString() says: "Formats a date in the date escape format yyyy-mm-dd".  So as 
far as I can tell, this means that, with the current implementation, the format yyyy-mm-dd will always be used, 
regardless of locale.  Is that right?  And is that _acceptable_??

2) If the answer to #1 is "Yes", then is it safe to replace the "writeString" call above with the following?

	SimpleDateFormat df = new java.text.SimpleDateFormat("yyyy-MM-dd");
	writer.writeString(df.format((java.sql.Date) val));

Or is there some reason why we need to avoid using SimpleDateFormat?

3) I noticed some DateFormat code in the engn, for Derby embedded.  Should Network Server just be using that somehow, 
instead of doing a different thing here?

Feedback/input?

Thanks,
Army