Easy selection of variables for filters / filtering in general
See original GitHub issueProblem statement
A recurring issue brought up in a test group is that setting up filters is hard. The reason is that, unlike demo data, all their datasets include units in parenthesis, often with unicode. This is very important for the user to know what data they are dealing with (compounded when metric and imperial units mix in a same dataset). That means backtick escaping, but it is also really hard and time consuming to type stuff like:
Max Operating Temp (°C)
Output Capacitance (μF)
Or things like subscripts or diacritics:
P₁₁
, P₁₂
, énergie cinétique
This is especially the case when you are live exploring in a meeting and get a question like “how does this perform under such and such conditions” and these are not in the legend.
Possible solution:
- drag a drop variables
- dataframe viewer supporting filtering at column header (like
qgrid
does)
A related issue is that once you have added the variable to the query, how do you figure out what values are valid / typical? This is especially problematic for categorical variables. Implementing solution 2 above would fix that. With 1, some form of contextual help would possibly work.
In the past, I’ve implemented something like 2 for a web based product and customers really liked that functionality.
Issue Analytics
- State:
- Created 3 years ago
- Reactions:1
- Comments:5 (5 by maintainers)
Top GitHub Comments
Regarding getting variable names into filters more easily, I think just paste is the best option. Drag and drop would be more work to implement and not really much more functional. I can implement copy/paste of single cells from DataFrame header/index cells, I’ll do that tonight won’t take long.
By the way feel free to add one of those datasets or similar to the
pandasgui.datasets
if it’s not private. Data with weird symbols like that would be good for test cases.