Hello,
We will prepare extensive document and videos to demonstrate this.
In short,
You can modify multiple rows (edit, delete or append) on client side and send all changes in a single batch request to the server
Navigation in grid won't change dataset cursor (This can be enabled using an option)
Editing cells doesn't send multiple events to the server as it happens in standard UniDBGrid
You can use database transactions to commit or rollback changes. This means that you can rollback all changes if an error occurs during posting multiple changes.
You can use non-datebase data sources to populate the grid. (You can write your own adapters to get data from any data source)
You can write custom event handlers to take full control on how data is posted to the underlying data source
You can work with closed data tables. (Open only when needed)
Unlike UniDBGrid, HyperGrid doesn't rely on "RecNo". Everything is handled using a KeyField which is a field in your table with unique key values. This key values will be used to locate rows when editing or deleting rows. Same key values are used when navigating in the grid.
Data adapters allow to centralize event handling for multiple HyperGrids using a single data adapter.
Modified rows won't be sent to the server until user decides to do so.