Java Doc for CachedResultSet.java in  » Database-JDBC-Connection-Pool » jTDS » net » sourceforge » jtds » jdbc » Java Source Code / Java DocumentationJava Source Code and Java Documentation

Notes:

1. For maximum performance use the scroll insensitive result set type.
2. As the result set is cached in memory this implementation is limited to small result sets.
3. Updateable or scroll sensitive result sets are limited to selects which reference one table only.
4. Scroll sensitive result sets must have primary keys.
5. Updates are optimistic. To guard against lost updates it is recommended that the table includes a timestamp column.
6. This class is a plug-in replacement for the MSCursorResultSet class which may be advantageous in certain applications as the scroll insensitive result set implemented here is much faster than the server side cursor.
7. Updateable result sets cannot be built from the output of stored procedures.
8. This implementation uses 'select ... for browse' to obtain the column meta data needed to generate update statements etc.
9. Named forward updateable cursors are also supported in which case positioned updates and deletes are used referencing a server side declared cursor.
10. Named forward read only declared cursors can have a larger fetch size specified allowing a cursor alternative to the default direct select method.

This result set will either be cached in memory or, if the cursor name is set, can be a forward only server side cursor.

This result set will either be cached in memory or, if the cursor name is set, can be a forward only server side cursor. This latter form of cursor can also support positioned updates.
Parameters:
  statement - the parent statement object
Parameters:
  sql - the SQL statement used to build the result set
Parameters:
  procName - an optional stored procedure name
Parameters:
  procedureParams - parameters for prepared statements
Parameters:
  resultSetType - the result set type eg scrollable
Parameters:
  concurrency - the result set concurrency eg updateable
exception:
  SQLException - if an error occurs

Sybase (and SQL 6.5) will automatically include any additional key and timestamp columns as hidden fields even if the user does not reference them in the select statement.

If the table is unkeyed but there is an identity column then this is promoted to a key.

Alternatively we can update, provided all the columns in the table row have been selected, by regarding all of them as keys.

SQL Server 7+ does not return the correct primary key meta data for temporary tables so the driver has to query the catalog to locate any keys. true if there is one table and it is keyed

If all the tables in the result set have primary keys then the result set row can be refreshed by refetching the individual table rows.
throws:
  SQLException - if an error occurs

Obviously very slow on large result sets but may be the only option if tables do not have keys.

Use to make result set read only once loaded.
Parameters:
  concurrency - the concurrency value egResultSet.CONCUR_READ_ONLY