DataTable shows old values after refresh

Dash Python
1

Hi everyone.

I’m displaying a real-time data with dash datatable. Data comes from a pulsar topic.

Flow is as follows:

I have a timer which is n_interval.
For each tick, I consume new data and draw datatable.
My problem is: old data (rows) is still on table even if new data does not contains that rows, they somehow still on table, so table is mixed with old and new data.

How can I ensure that when I display new data, old one is deleted?

Best regards.

Edit: My update table function is as follows:

@callback(
    [
        Output("table1", "data"),
        Output("table2", "data"),
        Output("last-update", "children"),
    ],
    [
        Input("timer", "n_intervals"),
        Input("status", "data"),
        Input("start-button", "n_clicks"),
    ],
    prevent_initial_call=True,
)
def update_tables(n, status, but):
    global consumer
    if (
        but != None
        and but > 0
        and status != None
        and status[0] == "Bağlantı VAR"
        and consumer != None
    ):
        msg = consumer.batch_receive()
        msg = msg[-1]
        # if message was published more than 5 seconds ago, ignore it
        while msg.publish_timestamp() < (time.time() * 1000) - 5000:
            consumer.acknowledge_cumulative(msg)
            msg = consumer.batch_receive()
            msg = msg[-1]

        data = msg.data().decode("utf-8")
        data = ast.literal_eval(data)
        # data = json.loads(data).values()
        ob = pd.DataFrame(data)
        ob = pl.from_dataframe(ob)
		# bunch of process with dataframe
        tsstr = ob["Time"].max().strftime("%H:%M:%S")
        tsstr = f"Last update: {tsstr}"

        global table1data
        global table2data
        global table1summary
        global table2summary

        table1data = ob.filter(pl.col("SomeColumn") == "A")
        table2data = ob.filter(pl.col("SomeColumn") == "B")

        table1summary = table1data.group_by("SomeGroupColumn").agg(
            [
                # aggregations
            ]
        )

        table2summary = table2data.group_by("SomeGroupColumn").agg(
            [
                # aggregations
            ]
        )

        return [
            table1summary.to_dicts(),
            table2summary.to_dicts(),
            tsstr,
        ]

    else:
        raise exceptions.PreventUpdate
2

Hi @mrtkp9993 welcome to the forums.

Could you add some information how you update the table?

3

I added code for my update tables callback.

4

I think that could be related to the use of glaobal variables. Please have a read here:

5

How can i consume a Redis pubsub or Pulsar topic without creating global variables? Creating a new client for each update is not viable.