Tabulator: incorrect data aggregation

[thuydotm]

ALL software version info

Software Version Info

panel 1.5.3

Description of expected behavior and the observed behavior

Observed behavior: Flat aggregators on Tabulator always give sums for all columns. Nested aggregators always return the first element in the group. Indexes displayed as NaNs. Data grouping works properly though.

Expected behaviors:

• aggregating correctly based on the specified aggregation methods
• indexes displayed appropriately.

Complete, minimal, self-contained example code that reproduces the issue

import pandas as pd
import panel as pn

df = pd.DataFrame([
    ('Germany', 2020, 9, 2.4, 'A'),
    ('Germany', 2021, 3, 7.3, 'C'),
    ('Germany', 2022, 6, 3.1, 'B'),
    ('UK', 2020, 5, 8.0, 'A'),
    ('UK', 2021, 1, 3.9, 'B'),
    ('UK', 2022, 9, 2.2, 'A')
], columns=['Country', 'Year', 'Int', 'Float', 'Str']).set_index(['Country', 'Year'])

nested_aggregators = {'Year': {'Int': 'sum', 'Float': 'mean'}}
flat_aggregators = {'Int': 'sum', 'Float': 'mean'}

flat_aggs_tabulator = pn.widgets.Tabulator(
    value=df, hierarchical=True, aggregators=flat_aggregators
)
nested_aggs_tabulator = pn.widgets.Tabulator(
    value=df, hierarchical=True, aggregators=nested_aggregators
)

pn.Accordion(
    ("Flat Aggs", flat_aggs_tabulator),
    ("Nested Aggs", nested_aggs_tabulator)
).servable()

Screenshots or screencasts of the bug in action