man-group
Enhancement request to allow auto-hiding of columns where all entries are None or Empty equivalent
My most common use case is for viewing data in dtale is displaying decoded Pandas data from database queries. The DB has a raw data column like a dictionary of fields:values that I unpack into individual columns. As columns are concatenated to the data frame, any field within a row that doesn't have applicable data is set to None type. So, after decoding, there are many None/NaN entries mixed throughout the data frame. Regardless of unpacked fields, each row will have a common 3 fields whos' combined combination describes the possible fields that apply to those rows. For example all rows with module: 3, parameter: 1, component: 1, would have non-None values for 3 applicable fields. When I've filtered view to a specific combination of these, it would be ideal if the columns that contain all None/Missing/Empty data could be automatically hidden and restored when removing the filter. When just using Jupyterlab or VSCODE, I typically do this with df.dropna(axis=1, how='all') to display a filtered copy. Since most of my dtale work is just viewing the different values of applicable fields for filtered rows, it would make it very useful to have the (some times like 50 columns) hidden so I can see only the columns that are applicable.
