Release Notes for GridSQL version 0.9
-------------------------------------

This software requires the Java Runtime Environment 5 (1.5) or greater. 
Please go to http://java.sun.com/javase/downloads/index.jsp 
and select Java Runtime Environment (JRE) 5.0. 

After installing, please make sure that your PATH environment
variable refers to the JRE's bin directory.

0.9 release notes
-----------------
The two new major features are Oracle compatiblity and support for 
the PostgreSQL protocol.

Oracle compatibility for Postgres Plus Advanced Server was added, 
including support for:
 - functions 
 - operators
 - data types
 - (+) outer join syntax (left joins only)
 - FROM DUAL syntax

Support for the PostgreSQL protocol was added.
This means there is no need to use the legacy GridSQL drivers;
users can connect to GridSQL using EnterpriseDB's or PostgreSQL's 
JDBC, ODBC and .NET drivers. Other drivers were not tested.
Users may also connect to GridSQL using psql or edb-psql, passing 
in the proper port with -p (default 6453).

To also facilitate moving to the PostgreSQL protocol, some 
additional SQL commands were added: CREATE DATABASE, DROP DATABSE,
START DATABASE, STOP DATABASE, and SHUTDOWN.

Support for a subset of the COPY command was also added. 
The gs-loader utility now internally makes use of the COPY protocol
and some of the options have been modified as a result.

In addition, various minor changes were made for better 
Postgres Plus Advanced Server integration.

Additional Changes:

- Made users global across GridSQL, instead of being specific to a database
- Added limited parameter support for PreparedStatements,
  and improved handling on server to avoid reparsing and replanning.
- Added support for EnterpriseDB SHOW commands
- Added support for ALTER TABLE RENAME COLUMN.
- Added support for ILIKE, SIMILAR TO, and ~
- Added support for current_user(), a synonym for user().
- Performance enhancements, including batches, ANALYZE, result combining, 
  UPDATE and DELETE

Known Issues
------------
In the current version, the performance of XDBLoader has degraded and is much 
slower than using COPY FROM. If performance is important to you and you don't 
need error handling for your loads, use COPY FROM.

COPY does not yet support the WITH CSV clause.

The xdbimpex utility has a bug with the -c option, commit interval. It is
recommended to use COPY FROM in the current release anyway.

Some forms of correlated subqueries in the HAVING clause currently do not
work.

ORDER BY on the boolean and interval data types is not sorted properly.


