DashTable only displaying rows where the columns values are different

So I have a df that I can display using data_table without any issues. I also compare between two columns of the df, and background color cells red of rows that are different. However, my issue is, how do I only display rows that are different just using the dashtable filtering? This is what I have so far in terms of displaying entire df using dashtable and only highlighting different rows based on Device1 and Device2 values.

        comptable = html.Div([html.H5(" "),
                              dash_table.DataTable(id='xmlreadtable',
                                data=comp_df.to_dict('rows'),
                                columns=[{'name': i, 'id': i} for i in comp_df.columns],
                                style_header={'backgroundColor': 'rgb(230, 230, 230)','fontWeight': 'bold'},
                                page_action='none',
                                export_format='csv',
                                style_table={'height': '770px', 'overflowX': 'scroll'},                              
                                style_cell={'textAlign':'left','minWidth': 40, 'maxWidth': 90, 'width': 90,'font_size': '16px','whiteSpace':'normal','height':'auto'},
                                style_data_conditional=[{
                                    'if': {'column_id': 'Device1', 
                                           'filter_query': '{Device1} != {Device2}'},
                                        'backgroundColor': '#FF0000',
                                        'color': 'white',
                                    },
                                    {
                                    'if': {'column_id': 'Device2', 
                                           'filter_query': '{Device1} != {Device2}'},
                                        'backgroundColor': '#FF0000',
                                        'color': 'white',
                                    },                                   
                                    ]
                                )
                             ])

I can create a new df which is a subset of the original df:

sub_df = df[(df['Device1'] != df['Device2'])]

But I would rather filter already displayed table. Also, will export_csv export only rows that are different or the entire df.