> 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.

# Import Parquet files from Azure Data Lake Storage

> Connect Mammoth to Parquet files in Azure Data Lake Storage with a storage account access key, pick columns, and save a dataset that refreshes.

The Parquet (Azure) connector reads Parquet files stored in **Azure Data Lake Storage** and imports them into Mammoth as a dataset that can refresh. Open it from **Data Library** > **New dataset** > **Connectors** and search for **Parquet**. Project Admins create the connection; Project Editors can start datasets from a connection the project already has. Which connectors you can add depends on your plan.

## What you can bring in from Parquet on Azure

- Parquet files in one Azure storage container. The container's folders (directory paths) group the files in **Tables**.
- Only the columns you tick.

This connector is for Parquet files on Azure. For Parquet files on Google Cloud Storage, use the [Google Cloud Storage connector](https://mammoth.io/docs/learn/get-data-in/connectors/google-cloud-storage/).

## Before you connect Parquet on Azure

You need:

- An Azure storage account that holds Parquet files.
- The storage account's **Access Key**, from **Azure Portal → Storage Account → Security + networking → Access keys**.
- The **Account** name and the **Container (File System)** name where the files are stored.

### Find your Azure storage access key

1. Open [Azure Portal](https://portal.azure.com) and go to your storage account.
2. Under **Security + networking**, select **Access keys**.
3. Copy **key1** (or key2).

## Connect Parquet on Azure to Mammoth

These steps follow the shared flow in [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/).

1. Go to the **Data Library** and click **New dataset**.
2. Choose **Connectors**.
3. Search for **Parquet** and click its tile. The connector workspace opens with the **Chat** showing.
4. In the **Chat**, Mammoth shows a card titled "Azure Data Lake Storage Credentials". Fill in the fields below.
5. Click **Connect**. Mammoth checks the details and saves the connection. If the check fails, the card shows the error.

![Parquet New Connection form titled Azure Data Lake Storage Credentials with Access Key, Account, and Container fields](https://mammoth.io/docs/images/20260615_112514_Screenshot2026-05-221750231.webp)



| Field | What to enter | Example |
| --- | --- | --- |
| **Access Key** | Azure Storage account access key (Key 1 or Key 2) | `abc123...` (a long string) |
| **Account** | Azure storage account name | `mydatalake` |
| **Container (File System)** | Blob container name where your Parquet files are stored | `analytics-data` |

## Pick Parquet tables and columns

1. In **Tables**, click a table to see its columns and sample rows.
2. Tick the columns you want. They gather in **Your table**.
3. Click **Preview 100 rows** to check the result.

If you're not sure which file holds what you need, ask in the **Chat**.

![Parquet New Dataset dialog with file list, SELECT query, data preview, Save table as, and Save as dataset button](https://mammoth.io/docs/images/20260615_112533_Screenshot2026-05-221723501.webp)



## Save Parquet data as a dataset

1. Click **Save as dataset**.
2. Enter a **Dataset name** and choose a folder under **Save in**.
3. Click **Save**. When the dataset is ready, click **Open dataset**.

The connector workspace treats Parquet as a file source, so it saves the data once and shows **Saved once. Scheduled refresh for files is coming.** There's no **Daily** or **Manual** choice when you save.

## Refresh Parquet data and choose an update method

To refresh the dataset, create a [Dataset Refresh](https://mammoth.io/docs/learn/automate/refresh-a-dataset/) automation for it.

After you save, the dataset's **Update method** setting offers these options:

| Update method | What it does |
| --- | --- |
| **Replace all data** | Rebuilds the dataset from the source on every refresh |
| **Add new rows only** | Keeps your rows and adds rows above the last value of a **Key column** |
| **Keep only the newest rows** | Empties the dataset, then brings in rows since the last refresh (needs a **Key column**) |

A key column is a numeric or date column whose values only go up, such as an ID or a date.

## Troubleshoot Parquet on Azure

| Symptom | Likely cause | Fix |
| --- | --- | --- |
| Authentication fails | The access key has a typo | Copy the key again. It's a long base64 string, and a single wrong character fails. |
| Authentication fails | The account name doesn't match | Enter the account name exactly; it's case-sensitive |
| Container not found | The container name doesn't match | Container names are case-sensitive in Azure. Confirm the container exists in Azure Portal under **Containers**. |

## Related

- [Connect a database or app](https://mammoth.io/docs/learn/get-data-in/connect-a-source/): the shared connector workspace steps.
- [Azure Blob Storage](https://mammoth.io/docs/learn/get-data-in/connectors/azure-blob-storage/): import CSV, Excel and other files from Azure with app credentials.
- [Google Cloud Storage](https://mammoth.io/docs/learn/get-data-in/connectors/google-cloud-storage/): import Parquet files from a GCS bucket.
- [Manage saved connections](https://mammoth.io/docs/learn/get-data-in/connections/): edit, reuse or delete this connection.
- [Refresh a dataset](https://mammoth.io/docs/learn/automate/refresh-a-dataset/): change when the dataset refreshes.

---
Source: https://mammoth.io/docs/learn/get-data-in/connectors/parquet.md · Updated: 2026-11-08