> AI agents: this is one page from Mammoth Analytics documentation. The index of all pages as Markdown is https://mammoth.io/docs/llms.txt. Append `.md` to any docs URL, or send `Accept: text/markdown`, to get Markdown.

# Refresh a connected dataset on a schedule

> Set up a Dataset Refresh automation to fetch the latest data for database, cloud app and From URL datasets on a schedule, in parallel or in a sequence.

A Dataset Refresh automation fetches the latest data for datasets from database and cloud app connectors, and for datasets that follow the link they came from, on a schedule or when you run it. Create one from **Automations** > **New automation** > **Dataset Refresh**, or choose **Set-up auto refresh** on a dataset. It is a plan feature, and your plan also sets how often it can run.

A dataset saved from a database or cloud app (Google Sheets, Salesforce, PostgreSQL and so on) holds the data from the moment it was saved. Dataset Refresh brings in the latest data from the source, following each dataset's update method.

Use it for daily CRM or ERP data pulls, Google Sheets you update often and query in Mammoth, or any source whose data changes often and should show up automatically.

## Which datasets you can refresh

The step that picks datasets says: "Only datasets connected to an API or database, or that follow the link they came from, can be set up for refresh."

| Can be refreshed | Can't be refreshed |
| --- | --- |
| Datasets from API and database connectors, including Google Sheets and Azure Blob Storage | Uploaded files |
| Datasets fetched [From URL](https://mammoth.io/docs/learn/get-data-in/import-from-a-url/) that follow their link | Datasets saved from file connectors: Dropbox, Google Drive, Google Cloud Storage, SFTP, SharePoint, OneDrive, Email (IMAP) |

To bring in new files from file sources, use [File Collection](https://mammoth.io/docs/learn/automate/collect-files/) with [Data Consolidation](https://mammoth.io/docs/learn/automate/consolidate/).

Uploaded files in the project are listed greyed out, marked "Uploaded files can't be refreshed". If no dataset qualifies, the step shows **No refreshable datasets** ("Connect a dataset to an API or database to make it available for refresh.").

### Plan rules for Dataset Refresh

Dataset Refresh is a plan feature (it needs the File Collection plan feature), and your plan sets how often it can run. If your plan has no scheduled refresh, the **Dataset Refresh** card shows a plan pill. If saving fails with "This feature is not included in your current plan." even though the card shows no plan pill, contact Mammoth support. See [What each plan decides](https://mammoth.io/docs/learn/automate/#what-each-plan-decides).

## Set up a Dataset Refresh automation

### Open the wizard on Dataset Refresh

Start from the Data Library header, a folder's or dataset's menu, the dataset panel's **Automation** section, or the **Automations** list. Each opens the **New automation** window.

![20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_0 (1).png](https://mammoth.io/docs/images/20261008_124420_20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_01.webp)

Under **Bring data in**, select **Dataset Refresh**. The wizard shows its steps (**Select datasets**, **Set Schedule**, **Review & save**) and opens the first one.

### Pick the datasets to refresh

1. In the **Select datasets to refresh** step, tick the datasets this automation should refresh. Each row shows the dataset's folder, its column and row counts, and its **Update Method**.
2. Use **Search datasets** to find one by name. Use **Clear all** to untick them all.
3. Check each ticked dataset's update method, and change it if needed. See [Choose what a refresh does with your data](#choose-what-a-refresh-does-with-your-data).
4. Click **Set Schedule**.

### Refresh several datasets in parallel or in a sequence

One automation can refresh several datasets on the same schedule. When you tick more than one, a **Refresh datasets** option appears:

| Option | What it does |
| --- | --- |
| **in parallel** | Refreshes all the datasets at once |
| **in a sequence** | Refreshes them one after another, in the numbered order shown next to each dataset. Drag a dataset by its handle to reorder |

A sequence keeps dependent datasets in the right order. For example, in a financial reporting refresh: transaction data first, then the customer master that depends on it, then product hierarchy reference data, then GL mappings.

### Set the refresh schedule

In the **Set Schedule** step, choose **On schedule** (you can still run it manually at any time) or **Manual only** (no automatic runs; it runs only when you trigger it manually or through the API).

For **On schedule**:

1. Pick **Day**, **Week**, **Month** or **Year**, and the number next to **Repeat every**.
2. Under **First run**, pick when the first refresh happens.
3. Optional: click **+ Add end date** and pick an **End date**.
4. Check the **Next runs** preview.

![Dataset refresh Set Schedule step: On schedule, Week, Repeat every 1 week on Monday, First run, + Add end date and the Next runs preview](https://mammoth.io/docs/images/20261102_dataset-refresh-set-schedule.webp)

The units, **Repeat every** values and date options are the same for every automation type. See [Schedules and triggers](https://mammoth.io/docs/learn/automate/schedules-and-triggers/). Without an end date, the automation runs until you suspend or delete it.

### Review and save the refresh

On **Review & save**, enter a descriptive name and an optional description. Review the summary (datasets, schedule and type), then click **Save automation**.

![20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_3 (1).png](https://mammoth.io/docs/images/20261008_124627_20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_31.webp)

## Choose what a refresh does with your data

Click a ticked dataset's update method to change it. The change applies to the dataset itself, not only to this automation: "This changes what happens the next time data is retrieved. Your current data stays as it is until then."

| Option | What it does |
| --- | --- |
| **Replace all data** | Rebuilds the dataset from scratch on every refresh, so it always matches the source |
| **Add new rows only** | Keeps your existing rows and adds the rows that appeared since the last refresh. Use it to build up history over time. Needs a **Key column** |
| **Keep only the newest rows** | Empties the dataset on each refresh and brings in only the rows that appeared since the last one. Everything older is removed. Needs a **Key column** |
| **Add to existing data** | Shown instead of **Add new rows only** for sources that can't use a key column: every refresh adds all the rows it brings in to the ones you already have |

A **Key column** is a column whose values only ever go up, like an ID or a date. Mammoth imports only rows above the last value it saw. Rows changed or deleted at the source afterwards are not updated in Mammoth. Without a key column, **Add new rows only** brings in a full copy of the source each time, so your rows are repeated, and **Keep only the newest rows** can't be saved.

The options offered depend on the connector. For more on each, see [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/).

![20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_1 (1).png](https://mammoth.io/docs/images/20261008_124523_20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_11.webp)

## Check what a refresh will do and when

The automation appears in the list with status **Active**. Its summary reads like *"Dataset products replaces the data every day at 6:00"*, or *"adds new rows"* when the update method adds rows. It runs at each scheduled time. You can also run it at any time with **Run now** from the detail panel or the three-dot menu.

Click the automation row to open the detail panel. The **Summary** section shows:

- **Refresh method**: the schedule, for example "On schedule, every day";
- **Next run**: when the next refresh is scheduled;
- **Ends**: the end date, or "Never";
- **Then sends**: the export steps a refresh sets off;
- **Dataset**: the datasets being refreshed.

### See which exports a refresh sets off (Then sends)

If Views of a refreshed dataset have export steps that run after their data updates, the row and the detail panel list them after **Then sends:**, one per export, with an icon for its destination. Paused or deleted export steps aren't listed. An export with a **Send at most** limit shows it after its name, as **(at most daily)** or **(at most weekly)**, so you can see it won't send on every refresh. A Scheduled Export's own list doesn't show these labels, because it sends on its own schedule.

## Edit a Dataset Refresh

Open the automation from the Automations list (or from a dataset's preview panel **Automation** section) and choose **Edit** from the three-dot menu. The edit window opens over the page you are on. There is no separate edit page.

![20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_4 (1).png](https://mammoth.io/docs/images/20261008_124659_20261008_050649_draft_01_learn_automation-creation-editing-moved-from-full-page-routes-to-in-place-modals-with-live-status-and-quick-actions-3_41.webp)

A banner at the top shows the status:

- **Active**: runs on its schedule. Click **Run now** for an immediate run, or **Suspend** to pause scheduled runs.
- **Suspended**: scheduled runs are paused. Click **Resume** to reactivate it.
- **Error**: the automation has an error.

In **DETAILS**, rename the automation or update its description. In **CONFIGURATION**, you can:

- click **Edit schedule** in the **Schedule** row to change the unit, **Repeat every**, weekdays, **First run** and **End date**;
- add datasets with **Add more**, or remove a dataset;
- with several datasets, drag to reorder them and choose **In parallel** or **In a sequence**;
- change each dataset's update method inline.

The automation **Type** is shown for reference but can't be changed. Click **Save changes** to apply your edits, or **Delete Automation** to move it to Trash. Changes take effect from the next scheduled run.

## Refresh a dataset by hand after a trial ends

When a trial ends and your plan doesn't include Dataset Refresh, the automation is suspended. Its row shows a **Refresh Manually** button ("This automation got suspended after trial. Refresh manually by retrieving latest data."). The same button is in the detail panel.

1. Click **Refresh Manually**. The dataset's settings open in the Data Library.
2. Retrieve the latest data once.

The automation itself stays suspended. If the dataset no longer exists, Mammoth shows "The dataset this automation refreshes is no longer available."

**Refresh Manually** appears only for a Dataset Refresh suspended because of the plan. An automation you suspended yourself shows **Resume** instead.

## Troubleshooting

| Symptom | Cause | Fix |
| --- | --- | --- |
| Status shows **Error** | The source connection expired or changed, or the source's structure changed in a way that doesn't fit the existing schema | Re-authenticate the connection in your Data Library. If the structure changed, check the source columns against the dataset |
| The automation runs but the data doesn't change | The source has not updated since the last refresh. This is expected: the automation pulls whatever the source holds | Check the source for new data, or use **Run now** after the source updates |
| Your rows are repeated after each refresh | **Add new rows only** has no **Key column**, so it brings in a full copy each time | Pick a key column whose values only go up, or use **Replace all data** |
| A dataset is greyed out with "Uploaded files can't be refreshed" | Uploaded files and file-connector datasets can't be refreshed | Use [File Collection](https://mammoth.io/docs/learn/automate/collect-files/) and [Data Consolidation](https://mammoth.io/docs/learn/automate/consolidate/) |
| Saving fails with "This feature is not included in your current plan." but the card shows no plan pill | A plan check refused the save | Contact Mammoth support |
| Saving fails with "Following frequency not allowed in your plan, please upgrade the plan" | The schedule is more frequent than your plan allows | Pick a coarser unit. See [Plan limits on how often an automation runs](https://mammoth.io/docs/learn/automate/schedules-and-triggers/#plan-limits-on-how-often-an-automation-runs) |

## Related

- [Schedules and triggers](https://mammoth.io/docs/learn/automate/schedules-and-triggers/): every schedule option, time zones and API triggers.
- [Manage automations](https://mammoth.io/docs/learn/automate/manage/): suspend, resume, run, edit or delete a refresh.
- [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/): create the connected datasets a refresh can update.
- [Import from a URL](https://mammoth.io/docs/learn/get-data-in/import-from-a-url/): datasets that follow their link can be refreshed.
- [Export on a schedule](https://mammoth.io/docs/learn/automate/scheduled-export/): send exports at fixed times instead of after each refresh.

---
Source: https://mammoth.io/docs/learn/automate/refresh-a-dataset.md · Updated: 2026-11-08