Aggregating/Summarizing a Model's Data

[wysota]

You have to provide your own QAbstractProxyModel implementation that performs such calculations. For performance reasons you probably want to calculate aggregates on request and cache them in the proxy model instead of recalculating them each time someone asks for data.

[Phlucious]

Thanks for the reply! I like the idea of a proxy model, but I don't understand the implementation. Wouldn't the proxy model have to create its own items to display the aggregates that are related to the original model items? My current understanding of proxy models is that they only filter/sort/re-format existing items.

[wysota]

Wouldn't the proxy model have to create its own items to display the aggregates that are related to the original model items?

Yes but that doesn't change anything. QSortFilterProxyModel also creates some internal data for the proxied model.

My current understanding of proxy models is that they only filter/sort/re-format existing items.

They can manipulate the model in any way.

If you don't feel comfortable with calling it a "proxy", you can implement it as a custom model completely independent of the model it observes. Just connect to appropriate signals of the original model and you're good to go.

[Phlucious]

Another thread helped me to understand that I've been approaching the model-view concept from an item-based approach and that had I never actually made the plunge to thinking about it in terms of models, so what you say now makes a lot more sense. Thanks for your help!