Re: FW: Question about the postgres resultset implementation

Wow, thanks! This will be very helpful to us. I can't believe the speed at which you've responded to this. I really
appreciateit. 

Are official releases of the JDBC driver tied to official releases of Postgres?

Phill


-----Original Message-----
From: Kris Jurka [mailto:books@ejurka.com]
Sent: Wed 10/13/2004 2:03 PM
To: Tornroth, Phill
Cc: pgsql-jdbc@postgresql.org
Subject: Re: [JDBC] FW: Question about the postgres resultset implementation



On Wed, 13 Oct 2004, Kris Jurka wrote:

> I believe Peter's suggestion was to use a hashmap to cache the
> string->index mapping so that findColumn would only need to do the
> equalsIgnoreCase on the first call of that string, meaning once per column
> instead of for every call.
>

I've committed a fix to cvs following this suggestion.  The new findColumn
method looks like this:

Kris Jurka

public int findColumn(String columnName) throws SQLException
{
        Integer index = (Integer)columnNameIndexMap.get(columnName);
        if (index != null) {
                return index.intValue();
        }

        final int flen = fields.length;
        for (int i = 0 ; i < flen; ++i) {
                if (fields[i].getColumnLabel().equalsIgnoreCase(columnName)) {
                        index = new Integer(i+1);
                        columnNameIndexMap.put(columnName, index);
                        return index.intValue();
                }
        }

        throw new PSQLException (GT.tr("The column name '{0}' was not found in this ResultSet.", columnName));
}