Transform on columnIndex with MultiIndex is broken in 1.4.0

[vitaliknet]

It appears after 1.4.0 transform refactoring the following code no longer works.

dg_ind.transform([{'type': 'filter', 'columnIndex': 1, 'operator': 'in',
'value': ['A', 'AA']
}])

dg_ind.data.index
MultiIndex([('2024-12-16 00:00:00+00:00', 'A'),
('2024-12-16 00:00:00+00:00', 'AA'),
...
('2024-12-16 00:00:00+00:00', 'ZZZ')],
names=[('date', ''), ('ticker', '')], length=8965)

Something like below does not work either.

dg_ind.transform([{'type': 'filter', 'column': ('ticker', ''), 'operator': 'in',
'value': ['A', 'AA']
}])

[ollyhensby]

Yes, I am also getting this issue in the latest release (v1.4.0) when trying to use transforms for multi-index columns.

Example code with multi-index:

import pandas as pd

# Sample data
data = {
    ('Sales', '2023'): [200, 220, 250],
    ('Sales', '2024'): [230, 210, 240],
}

# Create a DataFrame
df = pd.DataFrame(data)

# Set MultiIndex for columns
df.columns = pd.MultiIndex.from_tuples(df.columns, names=('Category', 'Year'))
grid = DataGrid(df)
display(grid)
grid.transform(
    [
        {
            "type": "filter",
            "columnIndex": 1,
            "operator": "in",
            "value": "200"
        }
    ]
)

Output:

^ We see that the table has not been transformed and is still showing all of the sample data.

---

Now if I apply the same transform BUT to a grid that is NOT multi-index:

import pandas as pd

# Sample data
data = {
    '2023': [200, 220, 250],
    '2024': [230, 210, 240],
}

# Create a DataFrame
df = pd.DataFrame(data)
grid = DataGrid(df)
grid.transform(
    [
        {
            "type": "filter",
            "columnIndex": 1,
            "operator": "in",
            "value": "200"
        }
    ]
)
display(grid)

Output:

^ This is now transformed as expected.