Table output: use selected row as a reaactive input

[corey-dawson]

In R Shiny, there are built in controls for selecting a row in a table output. The row index (of the row selected) can be used as an input to trigger other events. How can this functionality (user selected rows on a table) be used in Python Shiny version? Any example would be excellent. thanks in advance

Section 2.1.1 of this R DT manual 2.1.1 Row Selection

[bartverweire]

Hi,

you could use qgrid. An example is shown below. Note that you will need ipywidgets < 8.

"""
Requirements:
shiny
shinywidgets
ipywidgets==7.6.5
pandas
qgrid
"""
from shiny import *
from shinywidgets import *

import pandas as pd
import qgrid

app_ui = ui.page_fluid(
    # qgrid output widget
    output_widget("qgrid_output"),

    # text to put the feedback of row selection
    ui.output_text_verbatim("out_feedback")
)

def server(input, output, session):

    selected_rows = reactive.Value()

    @reactive.Calc
    def iris():
        """
        Load the iris data set
        """
        return pd.read_csv('https://raw.githubusercontent.com/mwaskom/seaborn-data/master/iris.csv')

    @output
    @render_widget
    def qgrid_output():
        """
        Render the qgrid table
        We also define a handler for the selection_changed event.
        """
        w = qgrid.show_grid(iris(), show_toolbar=False)
        # define the handler for a selection change
        # See https://qgrid.readthedocs.io/en/latest/index.html?highlight=selection_changed for info
        w.on("selection_changed", qgrid_select_handler)

        return w

    def qgrid_select_handler(event, widget):
        """
        The handler is a regular python function, not a reactive function.
        If we want to use the result in shiny, we need to update a reactive.Value
        :param event: event data
        :param widget: the widget on which the event occurred
        :return:
        """
        # Get the data frame that is displayed. This may differ from the original data frame
        # because of filtering, sorting, ...
        df = widget.get_changed_df()
        # Get the selected rows. This is an array of row indexes (after filtering, sorting, ...)
        sel = event["new"]

        # Set the value of the selected_rows reactive.Value to the data frame obtained by applying the index
        selected_rows.set(df.iloc[sel,:])

    @output
    @render.text
    def out_feedback():
        """
        Uses the reactive Value selected_rows, which was updated by the qgrid handler
        :return:
        """
        return selected_rows()

app = App(app_ui, server)

def main():
    run_app(app)

if __name__ == "__main__":
    main()

Kind Regards

Bart

[gshotwell]

You can do the same thing with shiny for python, here are the docs. We currently don't have programmatic selection, but we're starting work on that feature imminently.

https://shiny.posit.co/py/components/outputs/data-grid/#select-rows