QTreeView memory consumption

[eyeofhell]

Memory consumption is measured via task manager . Run app without connected model = 20mb, run app with connected model = 320mb O_O.
As far as i understand, using QSTandardItemModel will effectively reduce QTreeView to QTreeWidget - i will not be able to supply data dynamically as needed. For SQL queries that can be millions of records this will be a bit slow .

[tbscope]

Memory measurement via the task manager is like asking a baby the yield of a certain nuclear bomb :-)
It doesn't tell you the real information you want.

Nevertheless I believe you when your program starts using more and more memory. I also enjoyed using the list and tree widgets of windows.

Your understanding that using QStandardItemModel will reduce QTreeView to QTreeWidget isn't correct. QTreeWidget manages the items itself. Using a QTreeView together with QStandardItemModel changes nothing. It's a standard view and a standard model.

For SQL queries I suggest you look into one of the QSql...Model classes. Try it first with one of these and see if you still get big memory consumptions.

For a personal project I need to display thousands of items from a database in a treeview, and the memory consumption is minimal.

[eyeofhell]

It seems that if i replace QTreeView with QTableView in my example code memory is not consumed anymore O_O. So it seems that QTreeView is just not intended to be used for very huge amounts of data and QTableView must be used instead.

[dw]

For a personal project I need to display thousands of items from a database in a treeview, and the memory consumption is minimal.

Then maybe you could help me with some ideas.

I also have a big QTreeView and its model gets the display data from an sqlite database. So when the tree is drawed 2 queries are executed per row to get the data. This gives the view an annoying delay (for scrolling, expanding a branch etc.). Also when I have the database in the memory. If I use any styling on it that doubles the delay.

Didn't you have this problem? Should I initialize the model by reading everything from the database and building a hierarchical structure to the memory? Any other idea to improve performance?