Difference between revisions of "HSQLDB Integration"
From Apache OpenOffice Wiki
(→Desired Features) |
(→Desired Features) |
||
Line 18: | Line 18: | ||
* more tolerance when changing column definitions<br/>For non-professional users, it's quite common that a database schema is not defined once and then fixed, but instead evolves over time, as requirements evolve. HSQLDB currently is pretty restrictive in some areas, preventing some of such on-demand adjustments to the database.<br/>For instance, simply (?) changing the length of a column is prohibited as soon as the column is part of a relationship. | * more tolerance when changing column definitions<br/>For non-professional users, it's quite common that a database schema is not defined once and then fixed, but instead evolves over time, as requirements evolve. HSQLDB currently is pretty restrictive in some areas, preventing some of such on-demand adjustments to the database.<br/>For instance, simply (?) changing the length of a column is prohibited as soon as the column is part of a relationship. | ||
* updatable result sets<br/>This would allow updatable views (wouldn't it?) | * updatable result sets<br/>This would allow updatable views (wouldn't it?) | ||
+ | * column aliases as variables, see [http://www.openoffice.org/issues/show_bug.cgi?id=66837 issue 66837] |
Revision as of 07:10, 30 June 2006
This is a (more technically orientated) page about the integration of the HSQLDB engine[1] into OpenOffice.org.
Things To Do
- localize error messages (currently all hard-coded English)
- get rid of the "default_schema=true" parameter?
OpenOffice.org-specific settings
Connection settings which are currently necessary to connect to HSQLDB embedded in an OpenOffice.org database file
- storage_key: unique key which globally identifies the connection/storage
- fileaccess_class_name: Name of the class implementing the file access
- default_schema=true: lets HSQLDB behave as if it doesn't support schemas
Desired Features
- single-file
This would allow to get rid of the temporary files which are currently needed when working on the embedded HSQLDB, and thus remove some performance penalties - separation of the "auto increment" from the "primary key" concept. Currently, if you declare a column to be an auto increment column, it inevitable becomes the primary key
- editable views
Currently, if you create a view as simple asSELECT * FROM table
, this is not updateable, even iftable
itself is.
(side note: Hmm, have to check which part OOo plays here - perhaps it's OOo's fault in real?) - more tolerance when changing column definitions
For non-professional users, it's quite common that a database schema is not defined once and then fixed, but instead evolves over time, as requirements evolve. HSQLDB currently is pretty restrictive in some areas, preventing some of such on-demand adjustments to the database.
For instance, simply (?) changing the length of a column is prohibited as soon as the column is part of a relationship. - updatable result sets
This would allow updatable views (wouldn't it?) - column aliases as variables, see issue 66837