> 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.

# Export to Snowflake, BigQuery, PostgreSQL or SQL Server

> Send a Mammoth View to a dataset, a managed database, BigQuery, Snowflake, Databricks, PostgreSQL, MySQL, SQL Server, Redshift or Elasticsearch after every run.

Mammoth can write a View into a table after every pipeline run: another Mammoth dataset, a database Mammoth hosts for you, or your own database or warehouse. Click **Export** in the View toolbar, then pick a card on the **Mammoth Managed** or **External Destinations** tab. Send to Dataset, Snowflake and Databricks are on every plan; the other databases need a paid plan, and Mammoth-managed databases need Pro (see [Export destinations by plan](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/#export-destinations-by-plan)).

What every export shares (when it runs, send only new rows, run history, Auto-sync) is on [Export data after every pipeline run](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/).

## Send a View to another dataset

**Send to Dataset** pushes the current View's output into another Dataset, which can be in a different Project. Use it to pass processed results downstream in a multi-stage pipeline. Unlike **New dataset** in the **Save a copy** popover, it creates a persistent pipeline task that refreshes the Dataset on every run.

1. Click **Export**, open the **Mammoth Managed** tab and click **Send to Dataset**.
2. Choose **New Dataset** or **Existing Dataset**.
   - **New Dataset:** enter a name and pick a **Location** (a folder, which can be in another Project).
   - **Existing Dataset:** select a Dataset from the picker.
3. Click **Next**.
4. On the second step, set what is sent:
   - A filter condition, to send only matching rows.
   - **Include hidden columns**, to send columns hidden in the grid.
   - The column list, which maps each column to the destination. For an existing Dataset, **Hide already-mapped columns** helps you find the ones still to map.
   - **Each time this runs** (see below).
   - **Task Position**, in the form footer: click the pencil icon and tick **Keep at the end of the Pipeline** to keep the task at the end of the pipeline.
5. Click **Apply**.

| Each time this runs | What happens |
| --- | --- |
| **Keep this view's rows up to date** | Swaps the rows this View sent last time for its new ones. Rows other Views sent to the same Dataset stay |
| **Add a new copy of all rows** | Keeps every earlier run too, so totals grow on each run |

If other Views also send to the Dataset, an **Also fed by** note names them. If you switch an existing task from **Add a new copy of all rows** to **Keep this view's rows up to date**, the next run also removes the copies this View added on earlier runs.

Auto-sync for this task is switched from the step's options menu.

![Send to Dataset form in the pipeline panel, steps 1 and 2](https://mammoth.io/docs/images/20261009_003551_Export-send-to-dataset-steps-1-and-2.webp)


## Get a Mammoth-managed BigQuery or PostgreSQL database

Mammoth provisions a fully managed **Google BigQuery** or **PostgreSQL** instance on demand: no credentials to configure, no infrastructure to manage. Each appears as its own card on the **Mammoth Managed** tab, alongside Send to Dataset and Live link. They need the Pro plan.

1. Click **Export**, open the **Mammoth Managed** tab and click the **Google BigQuery** or **PostgreSQL** card.
2. Enter the destination table name (auto-filled from your Dataset and View names).
3. Click **Next: Generate Table and Copy Credentials**.
4. Mammoth creates the table, showing progress messages as it works, then opens the **Database Credentials** dialog.
5. Use those credentials to connect your BI tool (Tableau, Power BI, Looker) directly.

| Managed database | Credentials you get |
| --- | --- |
| **PostgreSQL** | Host, port, username, password, database and table, with a **Copy all credentials** button |
| **BigQuery** | The dataset and table, a **Service Account Email** to paste into tools such as Power BI, and a downloadable **Service Account JSON key** file. Keep the key secure |

![Mammoth Managed Google BigQuery export form with the table name field](https://mammoth.io/docs/images/20261009_003533_Export-form-mammoth-managed-bigquery.webp)

The managed database table is replaced in full on each pipeline execution (drop and recreate). Managed databases have no Keep-at-end option. A workspace that is over its managed-database limit gets an error when it creates a managed table.

## Supported external databases and warehouses

You provide the credentials; Mammoth delivers the data after each pipeline run.

| Destination | Default port or credential type | Typical use |
| --- | --- | --- |
| Google BigQuery | (JSON key file) | Cloud analytics, Looker integration |
| Snowflake | (password or key pair) | Cloud data warehouse |
| Databricks | (personal access token) | Lakehouse tables in Unity Catalog |
| Elasticsearch | 9243 | Search index, operational dashboards |
| SQL Server | 1433 | Microsoft ecosystem, Power BI |
| MySQL | 3306 | Web application backend |
| PostgreSQL | 5432 | General-purpose database, Tableau/Power BI |
| Amazon Redshift | 5439 | AWS data warehouse |

The footer of the export form shows the task's **Position** (default: end of pipeline) and **Update method**. Click the pencil icon to change them (see [Task position and update method](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/#task-position-and-update-method)).

## Allow Mammoth's IP address in your firewall

Whitelist Mammoth's IP address (`34.255.92.234`) in your database firewall before clicking **Validate Connection**. The required IP is shown at the top of the database, warehouse and Elasticsearch forms. See [Allow Mammoth through your firewall](https://mammoth.io/docs/learn/get-data-in/network-requirements/).

## Export to Google BigQuery

BigQuery uses a JSON service account credentials file, not a username and password.

1. Upload your BigQuery service account JSON file. If you have saved more than one connection, pick one from the **Connection** dropdown; an icon next to it deletes a saved connection.
2. Enter the **Table Name**.
3. Select your **Project Name** from the searchable dropdown (populated from your credentials). Each entry represents a project/dataset pair.
4. Choose the **Update method** in the form footer: **Replace**, **Combine** or **Merge**.
5. If using **Merge**, select the **Reference Column** values used as merge keys.
6. Optionally enable **Partitioning** (requires a DATE column in the View).

Mammoth also writes a reference table named `ref_<table>` alongside the data table in the BigQuery dataset.

### Partition a BigQuery table by date

Partitioning is available when the View contains DATE columns.

1. Tick the **Partitioning** checkbox and select a date column as the partition key.
2. Choose the partition granularity: **Day**, **Month**, or **Year**.

When editing an existing BigQuery export, the Partitioning checkbox is only enabled if the Update method is Replace. When creating a new export, Partitioning can be set up alongside Combine or Merge too.

![External Google BigQuery export form with Table Name, Project Name, partitioning and Replace, Combine or Merge update methods](https://mammoth.io/docs/images/20260422_170508_ExternaldestnationBigQuery1.webp)

## Export to PostgreSQL, MySQL, SQL Server or Amazon Redshift

These four databases share the same form.

| Field | What to enter |
| --- | --- |
| **Host** | Your database server hostname or IP address |
| **Port** | Pre-filled with the connector's default (3306 for MySQL, 5432 for PostgreSQL, 1433 for SQL Server, 5439 for Redshift) |
| **Username** | A database user with write permissions |
| **Password** | The database user's password |
| **Database** | The name of the target database |
| **Table name** (optional) | If left blank, Mammoth uses a default name derived from the Dataset and View |

1. Allow Mammoth's IP address in your firewall.
2. Fill in the fields, or pick a saved connection (below).
3. Choose **Replace**, **Combine** or **Merge** as the **Update method** in the form footer. With **Merge**, pick at least one **Reference Column**.
4. Click **Validate Connection**.
5. Click **Apply**.

**Validate Connection** confirms Mammoth can reach your database: it creates and then drops a test table. **Apply** only checks that the required fields are filled in. It isn't blocked by a failed or skipped validation, so validating first is the only way to catch bad credentials before the export runs.

**Include hidden columns:** a checkbox appears if your View has columns that are hidden in the grid. Enable it to include those columns in the export output.

Column names longer than 63 characters make the export fail, so shorten them first.

![External MySQL export form with an IP allowlist notice and Host, Port, Username, Password, Database and Table name fields](https://mammoth.io/docs/images/20260422_170634_MySQLexternaldestinationformshowingIPwhitelist1.webp)

### Reuse a saved database connection

PostgreSQL, MySQL, SQL Server, Amazon Redshift and SFTP exports remember the connection details you use. Once you have one saved, the form shows a **Connection** dropdown ("Select a saved connection") above the fields.

- Pick a saved connection to fill the form from it. A summary of the connection appears under the dropdown.
- Click **Enter details manually** to clear it and type new details.
- When you apply new details, Mammoth reuses an identical saved connection if there is one, so repeated exports to the same place don't pile up duplicates.
- If the saved connection was removed, the form says "This saved connection is no longer available. Enter the connection details below."
- If the list can't load, you see "Couldn't load saved connections. You can still enter details manually."

Google BigQuery has its own **Connection** dropdown for saved key files (see [Export to Google BigQuery](#export-to-google-bigquery)). To manage saved connections, see [Manage saved connections](https://mammoth.io/docs/learn/get-data-in/connections/).

## Export to Snowflake or Databricks

Snowflake and Databricks share one warehouse form. Table and column names are created exactly as written (quoted), so they are case-sensitive.

| Snowflake field | What to enter |
| --- | --- |
| **Authentication** | **Password** or **Key pair** |
| **Account identifier**, **Username** | Your Snowflake account and user |
| **Password**, or **Private key (PEM)** and **Private key passphrase (optional)** | The password, or the key pair |
| **Warehouse**, **Database**, **Schema** | Where the table goes |
| **Role (optional)**, **Table name (optional)** | The role to use and the table to write |

| Databricks field | What to enter |
| --- | --- |
| **Server hostname**, **HTTP path**, **Personal access token** | Your SQL warehouse connection |
| **Catalog**, **Schema** | Where the table goes |
| **Staging volume (optional)** | A Unity Catalog volume such as `/Volumes/main/sales/landing`. With it, data is bulk-loaded with COPY INTO; without it, rows are inserted in batches, which is much slower for large Views |
| **Table name (optional)** | The table to write |

Both forms show Mammoth's IP address to allow in your firewall, an **Include hidden columns** checkbox and **Validate Connection**. Choose **Replace**, **Combine** or **Merge** as the **Update method**; with **Merge**, pick at least one **Reference Column**.

## Export to Elasticsearch

| Field | What to enter |
| --- | --- |
| **Host** | Your Elasticsearch server hostname |
| **Port** | Default 9243 |
| **Username** | Elasticsearch username |
| **Password** | Elasticsearch password |
| **Batch Size** | Number of records per batch (default: 2,000) |
| **Index Name** | The Elasticsearch index to write to |
| **Connect Via** | HTTPS or HTTP (default: HTTPS) |

Whitelist Mammoth's IP address (`34.255.92.234`) in your Elasticsearch firewall, then click **Validate Connection** before **Apply**. Apply isn't gated on a successful validation, so validating first is the only way to catch bad credentials before the export runs.

**Include hidden columns** is available if your View has hidden columns. Elasticsearch is replaced on every run, so it can't send only new rows.

![External Elasticsearch export form with Host, Port, Username, Password, Batch Size, Index Name and Connect Via fields](https://mammoth.io/docs/images/20260422_170732_ElasticsearchformshowingHostPortUsername1.webp)

## Limits

| Item | Limit |
| --- | --- |
| Column name length (PostgreSQL, MySQL, SQL Server, Redshift) | 63 characters |
| Elasticsearch batch size | 2,000 records by default |
| Live links per View | 1 (see [Live links](https://mammoth.io/docs/learn/share-and-export/live-links/)) |

## Troubleshooting database exports

### Validate Connection fails

**Cause:** Mammoth's IP address (`34.255.92.234`, shown in the form) is not allowed in your database firewall, or the credentials lack permissions. Validate Connection checks for CREATE TABLE and DROP TABLE permissions.

**Fix:** Add the IP address to your firewall allowlist. Also grant INSERT permission, which the data delivery needs once the export runs.

### Schema conflict error

**Cause:** This applies to database and warehouse exports using the **Combine** or **Merge** update method, where the destination table persists between runs. The View's columns no longer match the destination table's columns, so the run fails.

**Fix:** Delete and recreate the destination, or adjust the View to match the original column structure.

### Merge fails

**Cause:** The Reference column is missing from the View or the destination table, or the names or data types don't match.

**Fix:** Confirm the Reference column exists in both with matching names and compatible data types.

### The export fails on long column names

**Cause:** A column name is longer than 63 characters.

**Fix:** Shorten the column name in the View (see [Rename, format and sort columns](https://mammoth.io/docs/learn/clean-and-shape/rename-format-sort/)).

For a destination that doesn't update, see [Troubleshooting exports](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/#troubleshooting-exports).

## Related

- [Export data after every pipeline run](https://mammoth.io/docs/learn/share-and-export/send-to-destinations/): plans, when it runs, run history and Auto-sync.
- [Send to files, apps and APIs](https://mammoth.io/docs/learn/share-and-export/send-to-files-and-apps/): SFTP, SharePoint, Google Sheets, Email, Tableau and REST API.
- [Live links](https://mammoth.io/docs/learn/share-and-export/live-links/): a CSV URL for Power BI and Excel.
- [Allow Mammoth through your firewall](https://mammoth.io/docs/learn/get-data-in/network-requirements/): IP addresses to allow.
- [Manage saved connections](https://mammoth.io/docs/learn/get-data-in/connections/): reuse database credentials.

---
Source: https://mammoth.io/docs/learn/share-and-export/send-to-a-database.md · Updated: 2026-11-08