QCompleter performance for large data models

[wysota]

Updating of the view is unlikely the problem, filtering the model is the bottleneck. Apparently upon each change in the text the whole model is refiltered. What you need to do is either track changes or provide a faster than linear way of filtering the model. Unfortunately the latter cannot be done by subclassing QCompleter so you will have to provide your own completer class for that.

[AlekseyK]

You can find FULL solution for such situation here: http://stackoverflow.com/a/33454284/630169