ALL software version info
panel 1.5.4
Windows, MacOS, Chrome, Edge
Description of expected behavior and the observed behavior
There appears to be a bug with the perspective pane automatically collapsing the hierarchy when the underlying data object is updated - this only seems to happen after clicking to expand or collapse a row. This makes it pretty annoying and difficult to use when a periodic callback is used to refresh the data.
I'd expect the depth of the hierarchy to be retained to whatever is currently visible.
Complete, minimal, self-contained example code that reproduces the issue
import pandas as pd
import numpy as np
import panel as pn
pn.extension('perspective')
def get_data(n=100):
return pd.DataFrame({'x': np.random.choice(['A', 'B'], size=n), 'y': np.random.choice(['C', 'D'], size=n), 'z': np.random.normal(size=n)})
table = pn.pane.Perspective(get_data(), group_by=['x', 'y'], sizing_mode='stretch_both')
def update_data():
table.object = get_data()
pn.state.add_periodic_callback(update_data)
pn.Row(table, sizing_mode='stretch_width', height=1000)

After collapsing "A", the rest of the levels will collapse and will get stuck

ALL software version info
panel 1.5.4
Windows, MacOS, Chrome, Edge
Description of expected behavior and the observed behavior
There appears to be a bug with the perspective pane automatically collapsing the hierarchy when the underlying data object is updated - this only seems to happen after clicking to expand or collapse a row. This makes it pretty annoying and difficult to use when a periodic callback is used to refresh the data.
I'd expect the depth of the hierarchy to be retained to whatever is currently visible.
Complete, minimal, self-contained example code that reproduces the issue
After collapsing "A", the rest of the levels will collapse and will get stuck