---
title: "Export Device Data"
canonical: "https://help.scopious.co.nz/space/KB/174719021/Export%20Device%20Data"
format: markdown
---
You can export device data for one or more devices in a device pool dating back to one year.

> Macro (toc)

## Export Data

1. Navigate to the **Device Pools** page using the navigation menu on the left side of the screen.
2. Click on the device pool that you would like to export device data for.
3. Click the three dots (…) context menu button on the top right of the screen and click **Export Device Data**.
4. Select the data export time range.
5. Select the devices to export data for.
6. Specify whether you want to export the full data set or the minimal data set. If you are not sure which to choose, see [Understanding Exported Data](https://scopious.atlassian.net/wiki/spaces/KB/pages/174719021/Export+Device+Data#Understanding-Exported-Data)
7. Click **Download**.

![Export device data.gif](media://c713bf23-52e1-4e87-8186-22612d00c263)

## Understanding Exported Data

When exporting the data, you can choose whether you want to export the full data set or the minimal data set. 

- Check this option if you want to use the exported data as a backup, or to import the result into another InfluxDB instance. The exported data will include all columns returned by our time series database.
- Uncheck this option if you want to use the export for data analysis. The exported data will only include the time, field and value columns.

### What will the full data set look like?

> ℹ️ The downloaded files will be in [InfluxDB v2 Annotated CSV](https://docs.influxdata.com/influxdb/v2/reference/syntax/annotated-csv/#csv-response-format) format and all timestamps will be in UTC.

The format of the full data set is the raw format returned from the time series database used to store the data. We don’t reformat or simplify the data so the file can be used directly as a backup to restore data in the future if needed.

The following information will help you understand how to interpret a file:

- Ignore the first 3 rows as these are annotation rows used by the database only.
- The 4th row contains the column headers. There are 3 columns that are of interest: `_time`, `_value`, `_field`
- The `_time` column defines the time the packet was received by the network server.
- The `_value` column defines the value contained within that packet for a particular field.
- The `_field` column defines the name of the field associated with that value.
  - The `_field` column contains all the fields reported by that device.
  - Each field is prefixed with `uplink_message_decoded_payload_` and so the words after that identify the field name.
  - The `_field` column contains a field named `uplink_message_frm_payload`. Ignore these rows as this field represents the raw data sent from the device before it gets decoded, and so is not human readable.

### What does the minimal data set look like?

The format of the minimal data set is simply a trimmed down version of the full data set that removes all columns that are not relevant to data analysis. Ignoring the first three annotation columns, it will only include:

- The `_time` column defines the time the packet was received by the network server.
- The `_value` column defines the value contained within that packet for a particular field.
- The `_field` column defines the name of the field associated with that value.
  - The `_field` column contains all the fields reported by that device.
  - Each field is prefixed with `uplink_message_decoded_payload_` and so the words after that identify the field name.
  - The `_field` column contains a field named `uplink_message_frm_payload`. Ignore these rows as this field represents the raw data sent from the device before it gets decoded, and so is not human readable.