# Mqtt.to() in tasks

**URL:** <https://community.influxdata.com/t/mqtt-to-in-tasks/25274>\
**Category:** InfluxDB 2\
**Tags:** flux\
**Created:** [June 2, 2022, 2:49pm UTC](https://community.influxdata.com/t/mqtt-to-in-tasks/25274 "2022-06-02T14:49:14Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Knaoser](https://avatars.discourse-cdn.com/v4/letter/k/3ab097/32.png) [@Knaoser](https://community.influxdata.com/u/Knaoser)\
**Post date:** [June 2, 2022, 2:49pm UTC](https://community.influxdata.com/t/mqtt-to-in-tasks/25274/1 "2022-06-02T14:49:14Z")

</div>

Hello,

I am using mqtt.to() to send data from my bucket to a broker, however the query doesn’t seem to work well,  
as you can see below :

 ![influxbug](https://us1.discourse-cdn.com/flex023/uploads/influxdata/original/2X/a/a6c3e304f503e9eb3eaa007ed6d76cbc1ec7f886.png)  
I get way more data from my broker than what I get from the database publish message, I have usually around 3 or 6 data per second while I have something like 50 data registered in the database per second.  
Here is the command used for the task, i can’t seem to get the \_time value in the msg payload as well,  
if someone knows how to extract the time ?

```auto
import "experimental/mqtt"

option task = {
    name: "Mqtt",
    every: 1s,
}

from(bucket: "test")
    |> range(start: -task.every)
    |> filter(fn: (r) => r["_measurement"] == "accel")
    |> filter(fn: (r) => r["_field"] == "Newtons")
    |> filter(fn: (r) => r["host"] == "ESP32")
    |> filter(fn: (r) => r["topic"] == "test")
    |> aggregateWindow(every: 1ms, fn: mean, createEmpty: false)
    |> yield(name: "mean")
    |> mqtt.to(
        broker: "tcp://host.docker.internal:1883",
        topic: "DB",
        clientid: "flux",
        timeColumn: "_time",
        qos: 0,
        valueColumns: ["_value"],
        tagColumns: ["test"],
    )

```

The (~50/per second) tables do appear in the normal querying system when I run the same query, so why doesn’t it work well in tasks ?

Thanks to anyone who can help

---

<div class="post-metadata">

**Author:** ![Knaoser](https://avatars.discourse-cdn.com/v4/letter/k/3ab097/32.png) [@Knaoser](https://community.influxdata.com/u/Knaoser)\
**Post date:** [June 2, 2022, 2:57pm UTC](https://community.influxdata.com/t/mqtt-to-in-tasks/25274/2 "2022-06-02T14:57:03Z")

</div>

> [@Knaoser](#):
>
> ```auto
> from(bucket: "test")
> |> range(start: -task.every)
> |> filter(fn: (r) => r["_measurement"] == "accel")
> |> filter(fn: (r) => r["_field"] == "Newtons")
> |> filter(fn: (r) => r["host"] == "ESP32")
> |> filter(fn: (r) => r["topic"] == "test")
> |> aggregateWindow(every: 1ms, fn: mean, createEmpty: false)
> |> yield(name: "mean")
> 
> ```

Ok I just found part of the problem, it seems that when I change the range start, I get more points per seconds ?? (Around |\> range(start: -10s) I get all the points)  
Do someone has an explanation and a solution on how to counter this ?  
I need to process data in timestamp close to real time processing and waiting 10 seconds for each frame of data is not the goal of my work