Qt models behind proxies and fetchMore()/canFetchMore() - c++

I am interested to know how would views behave in cases when the main data model with incremental data fetching is behind a proxy or chain of proxies.
How proxy which does item rearrangements like ungrouping proxy (example: http://lynxline.com/jongling-qt-models/) should implement support of fetchMore()/canFetchMore()?

Examining QSortFilterProxyModel's source allows me to conclude that:
QSortFilterProxyModel allows access to already fetched rows. Calling rowCount, data, etc. will behave like there is no more data in the source table.
QSortFilterProxyModel (and QAbstractProxyModel by default) routes canFetchMore and fetchMore calls to according methods of the source model.
It means that when you scroll down a view with a QSortFilterProxyModel populated from a dynamically fetched table, it will be dynamically populated with new data from source model. However you can't be sure that new items will be added to the end. Instead, when scrolling down the table, its data can completely change. New rows can be added anywhere depending on current sorting and filtering settings.
When no additional support of canFetchMore and fetchMore is given, all proxy models (including the linked UngroupProxyModel class) will behave like that by default.
QTableView behaves buggy displaying such models. For example, if the source model has rows -2; -3; 2; 3 (let ; be a row separator in my answer), the view without sorting would usually display something like -2; -3; 2; 3; empty row; empty row. Empty rows are displayed in the bottom of the table. When user scrolls down enough close to these items, the view requests fetching more data. However, when sorting is enabled, the view shows -3; -2; empty row; empty row; 2; 3, i.e. it moves empty invalid rows to the middle or to the top of the table. It doesn't affect functionality but looks awkward.
However, all aforementioned issues are implementation dependent. Refer to the code and documentation of the used model classes for more details. I think it's possible to implement proxy models that act however you like.
Also, it would be reasonable in some cases to fetch source model completely before displaying sorted data in the table. Proper sorting or joining is impossible while not all data has been fetched. If you fetched the source model data (e.g. while(model->canFetchMore()) { model->fetchMore(); }) then proxy models will behave exactly like the model is not dynamically populated.

Related

Display modified data from QAbstractListModel in QTableView

I have a QAbstractListModel that has a bunch of custom objects stored in it, and you can access the different fields of the custom objects in the model by specifying a role (if this is an improper use of Qt roles let me know because I must be confused). I want to display this data in a user friendly QTableView. I can get things displaying using a proxy model, but the issue is I don't want to display the raw values, I want to display specific data derived from the raw data. So for instance, I don't want a column for both ItemA.foo and ItemA.bar, I want to display just ItemA.foo - ItemA.bar in a single column. And to add to that, I want the automatic update functionality you get with models where if either ItemA.foo or ItemA.bar change, I want the difference column to automatically update and recalculate.
I would think that the way to do this would be to use some kind of table proxy model that listens to the source model, and then populates its own fields with the values derived from the source model and listens for dataChanged() signals from the source model. Then you plug this proxy model in to a QTableView. But to me this sounds like something that should be done in a view. Or is this something that should be done by the delegate? I could even go so far as to do these calculations in the base model itself and add roles specific to these values that should be displayed in the table, but that sounds like I am really overloading the responsibilities of the model.
TLDR: How do you manipulate data from a model in a QTableView? Should I do the data manipulation in the base model and then send that to the QTableView? Should I use a proxy model that manipulates the base data and sends it to the QTableView? Or am I completely misunderstanding something?
and you can access the different fields of the custom objects in the model by specifying a role
If you look at the documentation for Qt::ItemDataRole, you would see that Qt models should indeed provide different data for different roles but each role means some distinguished purpose of the data corresponding to the role. For example, the most commonly used role is probably Qt::DisplayRole as the data for this role defines the content displayed in the view e.g. it's the text in the table cell. If you are satisfied with all the other aspects of the view - font, background etc - you can just return empty QVariant for corresponding roles from your model, the view would figure out these details on its own. If you are using roles as a substitute for columns i.e. to return different pieces of conceptually the same data item, it is probably not the intended use of roles.
For the other part of your question - you can customize the appearance of data displayed in the view through the use of a delegate. For example, you could subclass QStyledItemDelegate, override its displayText method to display ItemA.foo - ItemA.bar instead of just ItemA.foo and then set this delegate into the column of your view corresponding to ItemA.foo via setItemDelegateForColumn. The tricky part here would be to detect changes in both ItemA.foo and ItemA.bar columns which would affect the text displayed in the delegate. A while back I implemented a dedicated class in one of my projects which listens to changes in some column of the original model and "reroutes" the change into another column through signal emitting. I did it to solve this very issue - to catch changes in what delegate should display although technically another model column is affected into which the delegate is not set.

Update of QTreeView without changing selection

I have a model that retrieves data from a table in a database from a certain SQL query, and shows the items in a QTreeView. The characteristics are:
the data comes from a table, but has an underlying tree structure (some rows are parents that have rows below them as children)
this tree structure is shown in the QTreeView
the children are selectable in the QTreeView (not so the parents)
the table in the database gets updated continuously
in the updates, a children can be added to any existing parent
periodically (with a QTimer) the QTreeView is updated with the contents of the table
Since the children are added at any time to any parent, the first silly approach when updating the QTreeView is clearing it all, and append all the rows again, in form of parent or children, to the QTreeView. This is a 0-order approximation, and it is indeed terrible inefficient. In particular, the following problems appear:
Any existing selection is gone
Any expanded parent showing its children is collapsed (unless ExpandAll is active)
The view is reset to show the very first row.
What is the best solution to this problem? I mean, the first solution I will try will be not to clear the QTreeView, but instead parse all the returned rows from the table, and check for each of them whether the corresponding item in the QTreeView exists, and add it if not. But I wonder if there is a trickiest solution to engage a given table in a database with a QTreeView (I know this exists for a QTableView, but then the tree structure is gone).
This thread mentions a general approach, but this might get tricky quickly, but I am not sure how this would work if the underlying model is changing constantly (i.e. the QModelIndex becoming invalid).
Worst case is that you will have to write your own mechanism to remember the selection before updating and then re-applying it.
I assume you use some model/view implementation? You could enhance your model with a safe selection handling, in case the example mentioned above does not work for you.
I guess this is the case for a self-answer.
As I presumed, after a careful analysis of what data is retrieved from the database, I had to do the following "upgrades" to the retrieval code:
I retrieve, along with the fields I want to show in the view, two identifiers, one for grouping rows and one for sorting items into groups
I also retrieve the internal record ID (an increasing integer number) from the table in the database, in order to ask only for new records in the next retrieval.
In the model population code, I added the following:
I first scan the initial records that may belong to existing groups in the model
When, in the scanning, I reach the last group in the model, this implies that the rest of retrieved records belong to new groups (remember the records are retrieved sorted such that items that belong to the same group are retrieved together)
Then start creating groups and adding items to each group, until we use all the records retrieved.
Finally, it is very important:
the use beginInsertRows() and endInsertRows() before and after inserting new items in the model
capture the sorting status of the view (with sortIndicatorSection() and sortIndicatorOrder()) and re-apply this sorting status after updating the model (with sortByColumn())
Doing that the current position and selection in the QTreeView receiving the model updates are preserved, and the items in the view are added and the view updated transparently for the user.

Which component of MV should transform raw data into user friendly format?

There is a structure that holds raw data (consider int) and a gui that should visualize this data. Gui part consists of a few models and and a few standard views (table and tree). The value should be presented either as a string or as a picture depend on user choice. There are some additional difficulties: in some cases the text is different for different elements - i.e., for the same column 0 normally means "Ok", but for a particular element it should be printed as "Yes", also depend on a user choice some columns should not print text for 0 values at all, living a cell empty and don't clutter the table; next, sometimes cell should show two values - e.g., current value and requested - something like "Yes -> No". This also applies to the case when user wants to see icons.
I see two options:
The model forms string text itself and store it somewhere (in the model item may be). The model also forms composite icon from base icons if it is needed. In that scenario I can use standard delegate and just use standard roles reimplementing QAbstractItemModel::data() method.
Reimplement delegates' paint() method (actually I need to reimplement delegates anyway because I need custom editors like QComboBox). The delegate can get any information it needs through the model and custom roles in paint() method and form the string or complex icon itself. As a suboption, I suppose, it's possible for a model to compose needed information itself, like returning std::tuple<int /*value*/, int /*requested value*/, int /*source element type*/, bool /*text or icon*/> in data() method for Qt::DisplayRole.
Update: I want add one thought about second option. Presumably delegates were introduced to handle the work of representing data. One of the consequences of that is the flexibility delegates can give. If you want to see a raw number in some column, you can just leave the default delegate, if you want some additional info presented in a cell you can set a special delegate for that specific column. And it seems you don't need even touch the model or something else.

Displaying big data using Qt Table View

I am not sure how to go about this big data problem. Instead of using a Table Widget, it is suggested to use a Table View for big data for speed. The data will be at most 200000 rows by 300 columns of strings. Also, it might have gaps in the structure. For the View, it is needed to define a Model. I was not able to do this yet and I wonder the following:
Will it really be faster to use a View if the user only wants replace some values in a column, and inspect the data? Is a QList a smart choice given speed needs, or are there better options? And how to implement this as a Model?
A QTableWidget uses its own Model as backend (container for its data). One can't (or it is not intented that one should) customize a QTableWidget's model. Each cell in a QTableWidget is represented by a QTableWidgetItem. They are written to solve general problems and are not that fast.
When one decides to use a QTableView with it's own model deriving from QAbstractTableModel as backend, one can speed things up allot. I've found that one of the things that took the most processing time was the size calculation of rows. I overloaded sizeHintForRow in my table like this:
int MyTable::sizeHintForRow ( int row ) const
{
//All rows have the same height in MyTable, hence we use the height of the
// first row for performance reasons.
if( cachedRowSizeHint_ == -1 )
{
cachedRowSizeHint_ =
model()->headerData( 0, Qt::Vertical, Qt::SizeHintRole ).toSize().height();
}
return cachedRowSizeHint_;
}
Implementing a simple table model is quite easy (Qt Help files are adequate...). One can start by providing the data for the minimum roles (Edit and Display). TableModel data can always be modified by using QIdentityProxyModels where necessary. If your models don't work, post your code. Remember to not return invalid data for roles e.g:
QVariant IO_PluginMngr::data(const QModelIndex & index, int role) const
{
if( index.isValid() )
{
if( role == Qt::DisplayRole)
{
return sequence_[index.row()][index.column()];
}
}
return QVariant();
}
Returning for example erroneous data for especially SizeHintRole will cause havoc in the views.
I've noticed that one should, when implementing models, be diligent in calling begin/endInsertRows when appropriate, otherwise persistent model indexes don't work.
We've written more complicated tables where the data is modified in threads in the background, and swapped out prior to updating the table. One of the areas in which one could have speed improvement, is to never order the data in your model, but rather use an ordered reference sequence that references the real (unordered) data. We have tables that have about a thousand rows and 50 columns that get updated once a second (with many views), and the processor our processor is sitting a about 20%, but the greatest speed improvement was the caching of the sizeHint.

Supporting multi-add/delete (and undo/redo) with a QAbstractItemModel (C++)

Greetings,
I've been writing some nasty code to support the undo/redo of deletion of an arbitrary set of objects from my model. I feel like I'm going about this correctly, as all the other mutators (adding/copy-pasting) are subsets of this functionality.
The code is nastier than it needs to me, mostly because the only way to mutate the model involves calling beginInsertRows/beginRemoveRows and removing the rows in a range (just doing 1 row at a time, no need to optimize "neighbors" into a single call yet)
The problem with beginInsertRows/beginRemoveRows is that removal of a row could affect another QModelIndex (say, one cached in a list). For instance:
ParentObj
->ChildObj1
->ChildObj2
->ChildObj3
Say I select ChildObj1 and ChildObj3 and delete them, if I remove ChildObj1 first I've changed ChildObj3's QModelIndex (row is now different). Similar issues occur if I delete a parent object (but I've fixed this by "pruning" children from the list of objects).
Here are the ways I've thought of working around this interface limitation, but I thought I'd ask for a better one before forging ahead:
Move "backwards", assuming a provided list of QModelIndices is orderered from top to bottom just go from bottom up. This really requires sorting to be reliable, and the sort would probably be something naive and slow (maybe there's a smart way of sorting a collection of QModelIndexes? Or does QItemSelectionModel provide good (ordered) lists?)
Update other QModelIndeces each time an object is removed/added (can't think of a non-naive solution, search the list, get new QModelIndeces where needed)
Since updating the actual data is easy, just update the data and rebuild the model. This seems grotesque, and I can imagine it getting quite slow with large sets of data.
Those are the ideas I've got currently. I'm working on option 1 right now.
Regards,
Dan O
Think of beginRemoveRows/endRemoveRows, etc. as methods to ask the QAbstractItemModel base class to fix up your persistent model indexes for you instead of just a way of updating views, and try not to confuse the QAbstractItemModel base class in its work on those indexes. Check out http://labs.trolltech.com/page/Projects/Itemview/Modeltest to exercise your model and see if you are keeping the QAbstractItemModel base class happy.
Where QPersistentModelIndex does not help is if you want to keep your undo/redo data outside of the model. I built a model that is heavily edited, and I did not want to try keeping everything in the model. I store the undo/redo data on the undo stack. The problem is that if you edit a column, storing the persistent index of that column on the undo stack, and then delete the row holding that column, the column's persistent index becomes invalid.
What I do is keep both a persistent model index, and a "historical" regular QModelIndex. When it's time to undo/redo, I check if the persistent index has become invalid. If it has, I pass the historical QModelIndex to a special method of my model to ask it to recreate the index, based on the row, column, and internalPointer. Since all of my edits are on the undo stack, by the time I've backed up to that column edit on the undo stack, the row is sure to be there in the model. I keep enough state in the internalPointer to recreate the original index.
I would consider using a "all-data" model and a filter proxy model with the data model. Data would only be added to the all-data model, and never removed. That way, you could store your undo/redo information with references to that model. (May I suggest QPersistentModelIndex?). Your data model could also track what should be shown, somehow. The filter model would then only return information for the items that should be shown at a given time.