whish 2016

[stonebig]

By order of (complex) whish:

• match/do-better-than user simplicity of "excel slicers",
• scale over 1 000 000 rows,
• update goals and/or dates on your github milestones,
• add "ttk" in your setup.py keywords.

[dmnfarrell]

Can you explain what you mean by the first two in more detail? The table will scale up to >1 million rows but it will take time to load a large dataframe or csv file. A progress bar is needed for that.
I can add ttk" to the setup.py file.

[stonebig]

The first two are Ease-of-use problems I encounter when browsing (my sort of) Dataframes with the Python stack.

• no (Excel-like) Slicers:
  • a slicer is a multi-selectable list of uniques values present in a given column:
  • sorted by "shown" in current sub-selection of dataframe, then "hidden"
  • sub-sorted in alphabetical order
  • several Slicers can be shown to user to sub-select a given dataframe,
  • when you click on one slicer choice, the other slicers are re-draw to show remaining 'visible' values first,
  • for a Dataframe of sales of a Portfolio of Product with a "Unit Weight" slicer, the user can select his own definition of "small units" and see immediately the effect on the Dataframe (in its 'resume' form, a Pivot Table or a graphic in Excel case)
• browsing over 1 000 000+ rows
  • 400 000 rows is a limit at which Excel becomes unusable (too slow),
  • in my personnal experiment, loading a big Dataframe in a Graphical Python Library becomes unusable at much lower than 400 000 rows, so managing 1 000 000 rows mean being able not to load them all but only a window (a complex machinery).

The "ttk" tag is to help increase your visibility on Pypi.