DatabaseIO().write(database, file) does not write index information

[Thorbjørn Ravn Andersen <th...@gmail.com>]

I wish to use ddlutils to compare two schemas to see if they are 
identical, and if not, where.  I am starting with Oracle but will 
include DB2 plus some java based database for testing very soon.

I have the following snippet:

    private static Database getDatabase(String schema, String url,
            String username, String password, String name, Properties p)
            throws SQLException {
        Connection jdbcConnection = DriverManager.getConnection(url, 
username,
                password);

        JdbcModelReader r = new JdbcModelReader();
        String thiscatalog = null;
        String thisschema = schema;
        Database d = r.getDatabase(jdbcConnection, thiscatalog, thisschema,
                new String[] {});
        Table[] t = d.getTables();
        System.out.println(Arrays.asList(t));

        return d;
    }

which I call like this:

            d = getDatabase(schema, url, username, password, name, p);
            new DatabaseIO().write(d, "vmtra03-axs.xml");

which returns the tables I expect (after upgrading to the 10g JDBC 
driver).  It does not, however, return index information on the tables, 
which is needed too.  I looked in the code, and saw that apparently this 
requires additional invocations to get loaded, but wanted to ask on the 
status on the list first.

I also expect to have to deal with sequences, but have not gotten that 
far yet.

Could a knowledgeable person let me know how to include index 
information to what I have?

Best Regards,
-- 
  Thorbjørn

Re: DatabaseIO().write(database, file) does not write index information

[Thomas Dudziak <to...@gmail.com>]

On 11/4/05, Thorbjørn Ravn Andersen <th...@gmail.com> wrote:
> I wish to use ddlutils to compare two schemas to see if they are
> identical, and if not, where.  I am starting with Oracle but will
> include DB2 plus some java based database for testing very soon.

<snip>

> It does not, however, return index information on the tables,
> which is needed too.  I looked in the code, and saw that apparently this
> requires additional invocations to get loaded, but wanted to ask on the
> status on the list first.

Unfortunately, the database -> model part is not complete yet. But you
add an issue to JIRA stating that index reading does not work with
Oracle.

> I also expect to have to deal with sequences, but have not gotten that
> far yet.

I think this doing this in a generic way is only possible for a very
limited set of sequences (namely auto-increment sequences).

Tom