BUG- When a table is filtered, `on_action()` does not pass the correct row number of the original DataFrame

[Dr-Irv]

Description If a user filters a table, and an on_action() callback is set on the table, then the payload["index"] contains the row number of the filtered table, but there is no way to know what real row that is in the original DataFrame.

We are pretty sure that the same error occurs with on_edit() as well.

How to reproduce

from typing import Any

import pandas as pd

from taipy.gui import Gui, Markdown, State

table1 = pd.DataFrame({"x": [1, 2, 3], "y": [4, 5, 6]})

indexval = -1
xval = "-1"
yval = "-1"

page = Markdown(
    """

<|{table1}|table|filter=True|rebuild=True|width=fit-content|on_action=on_action|>

<|Indexval|text|> <|{indexval}|text|>

<|xval|text|> <|{xval}|text|>

<|yval|text|> <|{yval}|text|>

"""
)

def on_action(state: State, var_name: str, action: str, payload: dict[str, Any]):
    indexval = payload["index"]
    xval = state.table1.iloc[indexval]["x"]
    state.xval = str(xval)
    yval = state.table1.iloc[indexval]["y"]
    state.yval = str(yval)
    state.indexval = indexval

if __name__ == "__main__":
    gui = Gui(page=page)
    gui.run(title="Filter Action")

Sequence of events:

• Filter the table on "x" "equals" "3"
• That will leave one row. Choose that row
• The action updates the index values and the corresponding row of the DataFrame, which is not the filtered value

Expected behavior The row number should be of the filtered DataFrame, or there should be some way to know what rows remain after filtering.

Runtime environment Please specify relevant indications.

• OS: Windows 10 2H22
• Browser: Chrome 114.0.5735.134
• taipy 2.3.1

[FlorianJacta]

Good find! Using the index in the payload seems impossible as it corresponds to the filtered data.