Dynamic column name in `add_filter` method of tabulator

I tried to modify the tabulator filtering example by replacing the column string argument with the widget (to be able to filter value ranges of a dynamically selected column):

import pandas as pd
import panel as pn
mixed_df = pd._testing.makeMixedDataFrame()
filter_table = pn.widgets.Tabulator(mixed_df)
slider = pn.widgets.RangeSlider(start=0, end=5, name='A Filter')
select = pn.widgets.Select(value = 'A', options=mixed_df.select_dtypes(include=["float64","int64"]).columns.to_list(), name='C Filter')

    pn.Column(slider, select),

Unfortunately it looks like this is not implemented. Is there any way to achieve this functionality? I tried to do it with pn.bind or @pn.depends but wasn´t successful. I´m not very experienced and might have a knot in my brain :smiley:.
I finally could do it by modifying add_filter method directly in tables.py and modifying the col_name to column = df[col_name.owner.value] in _filter_dataframe (it´s not very clean programming and I´m not sure if I can do this in a general way for everyone but it solves my problem for the moment and might help for future implementation, so, here is the code:)

    def add_filter(self, filter, column=None):
            #######MODIFIED PM###################        
        if isinstance(filter, (tuple, list, set)) or np.isscalar(filter):
            deps = []
        elif isinstance(filter, (FunctionType, MethodType)):
            deps = list(filter._dinfo['kw'].values()) if hasattr(filter, '_dinfo') else []
            filter = param_value_if_widget(filter)
            column = param_value_if_widget(column)
            if not isinstance(filter, param.Parameter):
                raise ValueError(f'{type(self).__name__} filter must be '
                                 'a constant value, parameter, widget '
                                 'or function.')
            elif column is None:
                raise ValueError('When filtering with a parameter or '
                                 'widget, a column to filter on must be '
            deps = [filter]
            deps2 = [column]
        for dep in deps:
            dep.owner.param.watch(self._update_cds, dep.name)

        for dep2 in deps2:
            dep2.owner.param.watch(self._update_cds, dep2.name)            
        self._filters.append((column, filter))
            #######MODIFIED PM###################

On the other hand I think the client side filtering is very attractive, as the input fields are directly above the column. This works great for strings or “>” “<” filtering but not for ranges.
Here it would be great to find a way to implement the min/max input field from:


Because you can pass a function to add_filter, and in particular a bound function (bound with pn.bind), it’s really flexible and you can achieve the filtering on a given column you want:

import pandas as pd
import panel as pn


mixed_df = pd._testing.makeMixedDataFrame()
table = pn.widgets.Tabulator(mixed_df)

slider = pn.widgets.RangeSlider(start=0, end=5, name='A Filter')
select = pn.widgets.Select(
    name='C Filter'

def filter_table(df, column, rng):
    return df[(rng[0] <= df[column]) & (df[column] <= rng[1])]

table.add_filter(pn.bind(filter_table, column=select, rng=slider))

    pn.Column(slider, select),
1 Like

As for the min-max header filter of the Tabulator JS library, if you have a look at the source code of the example (Tabulator) you’ll see that it’s a very customized one, created by hand, so it’s not just a parameter Panel would have to set. I agree it looks awesome though, it’d probably be first a feature request to Tabulator itself before being able to have it in Panel.

1 Like

Thank you very much!
The thing that confused me was the return value of the pn.bind being a dataframe and not (filter,column). But obviously it works like the function based filtering example. Sorry I couldn’t put 1and1 together.

As for min/max in Tabulator I posted a feature request,let’s see .

1 Like