I've always hesitated to use QPersistentModelIndex when deriving my own model classes. I don't really know why. Maybe it's the lack of examples that use them or the skimpy "Detailed Description" section in the documentation. But I guess it's mostly that I felt they could/would cause a lot of overhead maintaining them. My guess is that they would have little overhead unless items are being added/removed/moved. Is this true?
Would I see a performance hit using persistent indices instead of rolling my own lookup functionality? I have ~5 model classes possibly containing thousands of indices each most of which are displaying real-time data. These models are then used as inputs into various QSortFilterProxyModels.
The following article discusses the 3 ways of handling the custom data mapping but it doesn't go into any detail regarding QPersistentModelIndex.
http://trolltech.com/developer/faqs/...tentModelIndex
Any insights?
Thanks,
RobF
Bookmarks