Use unique indices to identify datasets for appending #449
Conversation
|
Thanks @bkloeckl! The current changes lack the ability to sort datasets (via drag and drop) in the Source and Destination lists though, which I think we should not remove. I would also keep it simpler, instead of using a table, I'd stick to a list, but just add the index to the label. This could also be done for the sidebar then. I'm not sure how and if such a label can be formatted nicely, e.g. it would be nice to have
Something like this: 
If this requires a table (which might very well be the case), I'm sure that the style can be adapted so that it doesn't contain a header row or any borders. |
…functions to include original indices used for identifying the data
cbrnr
force-pushed
the
duplicate_names
branch
from
7a2b338 to
211dabf
Compare
|
I found an issue with the current implementation: Screen.Recording.2024-11-29.at.09.13.36.mov |
|
Another thing that could be improved is that currently, dragging one or more items from the source table to the destination table (or vice versa) shows the drop hint (the black line where the item would be inserted), but unfortunately dropping to the other table does not work. I think it would be very nice if that worked (then users could operate the dialog only via drag and drop and would not even need the arrow button, which is still a useful backup/alternative). |
|
Finally, I was wondering if you really have to implement the drag and drop on that level, or if you could maybe just connect the |
For the sidebar, the data and its order are managed by the model and then visually updated. The append-dialog has to handle the reordering on its own. This implementation is also the base for the latest update, where drag&drop works across the source and destination tables. |
|
Thanks @bkloeckl! |
* Change the append dialog appearance, get_compatibles and append_data functions to include original indices used for identifying the data * Add changelog entry * Ruff formating * ruff formating * ruff formating imports * adapted tests * adjust indices when needed after auto_duplicate() * adjust test * delete wrong test_model, commit correct one * change append dialog styling and add drag-drop functionality * table styling changes * align correction * Fix changelog * Improve readability * Minor fix * Fix style * fix move-bug by sorting rows * drag&drop between source and destiantion table now possible * Style * ruff error? * Fix style * Reduce redundancy * Use global constant for row height * Remove viewport update (not necessary) * Remove from paintEvent * Simplify (no set necessary) * Restore horizontal line * Remove leftover code * Remove another set
Change the append dialog, get_compatibles and append_data functions to include original and unique indices used for identifying the data.