gt_merge_stack doesn't seem work well with other parts of gt

[avamsi]

I'm trying to summarize some percentages along with counts providing context.

I considered the following options to merge the 2 (in my case 2 * N) columns:

1. Wrap them in a custom object (with it's own repr) but that wouldn't work with GT(?).
2. Serialize them as a string (with the same formatting I'd use in repr above) but that wouldn't work with data_color (though I suppose I could write my own helper using tab_style or something).
3. Use gt_merge_stack (just came across it), which seems to be working well with data_color (great!) but seems to be breaking at least autocolor_text and fmt_number.

Are there any other solutions to what I'm trying to do? If not, do you think this is a reasonable feature request?

Example:

(
    GT(pl.DataFrame(dict(a=[1 / 2, 2 / 3, 3 / 4, 4 / 5], b=[4] * 4)))
    .data_color(c.a, palette="GnBu")
)
(
    GT(pl.DataFrame(dict(a=[1 / 2, 2 / 3, 3 / 4, 4 / 5], b=[4] * 4)))
    .pipe(gte.gt_merge_stack, c.a, c.b)
    .data_color(c.a, palette="GnBu")
)

^ Notice the difference in precision and text colors (I do realize gt_merge_stack has it's own colors and is likely overriding colors from data_color, but I hope they can be made to work well with each other).

[juleswg23]

Hi @avamsi, you make a good point that is really a problem with the entirety of the gt-extras approach.

Nearly every function is a wrapper under the hood that calls GT.fmt(), but fmt() calls don't stack in great-tables. So since the precision adjustment relies on a fmt() call, it gets overridden by the fmt() that happens in gt_merge_stack(). I'm not entirely sure what we can do about this beyond opening an issue in Great Tables.

As for the colors, you are right. This is probably a solvable problem, but for now you do have the ability to change the color_main= and color_secondary= params.