Set up a Mammoth export and pick a destination
An export delivers a View's data to a destination every time the pipeline runs: a database or warehouse, a file store, an app, an API or another dataset. Click Export in the View toolbar, pick a destination and click Apply. The export becomes a task pinned to the end of your pipeline; which destinations you can use depends on your plan.
Set up an export
- In the View toolbar, click Export. The Export menu, titled with the View's name, opens on the External Destinations tab.
- Pick a destination card, or switch to the Mammoth Managed tab for Send to Dataset, Live link or a managed database.
- Fill in the export form in the pipeline panel. A summary row with a pencil icon lets you switch to another category or connector.
- Click Apply. The export runs straight away and after every pipeline run.
| Tab | Who manages it | Use when |
|---|---|---|
| External Destinations | You provide credentials or connect an account | You have your own database, warehouse, server or app |
| Mammoth Managed | Mammoth provisions everything | You want zero infrastructure setup |

Which export destination to use
| Situation | Best option |
|---|---|
| Colleagues need the file in their inbox after each update | Export → Email |
| A team works from a shared spreadsheet | Export → Google Sheets (Export to files and apps) |
| Data belongs in your warehouse | Export → Snowflake, Databricks or another database (Send data to a database or warehouse) |
| A BI tool needs to query Mammoth data via URL | Export → Live link |
| Tableau or Power BI needs automatic data delivery | Export → a managed or external database, or Tableau Server/Cloud |
| Push results into another Mammoth Dataset | Export → Send to Dataset |
| An API Mammoth has no card for | Export → REST API Export or Somewhere else… (Export to an API) |
| A destination must get data at fixed times, even when nothing changed | Export, then an Export on a schedule automation |
| An email should go out at most once a day or week | Export → Email, with Send at most |
| Someone needs the data once | Not an export: download a copy |
Export destinations by plan
A card shows a Trial pill when your plan doesn't include that destination. Hover the pill to see which plans include it ("Included in" and the plan names). Clicking a gated card opens the upgrade prompt instead of the form.
| Destination | Plan |
|---|---|
| Send to Dataset, Snowflake, Databricks, Google Sheets, Email | Every plan, no Trial pill |
| PostgreSQL, MySQL, SQL Server, Amazon Redshift, Google BigQuery, Elasticsearch | Paid plans (database exports) |
| SFTP, SharePoint, OneDrive, Azure Blob Storage | Paid plans (file exports) |
| Live link, Tableau Server/Cloud, REST API Export | Team and Pro |
| Mammoth-managed Google BigQuery and PostgreSQL | Pro |
| Somewhere else… (AI custom destination) | Available on Enterprise plans (it opens the AI Agent) |
A workspace that is over its managed-database limit gets an error when it creates a managed table.
Task position and update method
The form footer shows a read-only Task Position line and, where the destination has one, an Update Method line. Click the pencil icon to edit them.
Keep the export at the end of the pipeline
By default the export task sits at the end of the pipeline, so it runs after all transformation tasks, and Mammoth keeps it there. Change it only if you need a specific position (rare). Live link and the managed databases have no position option.
Choose how data is written: Replace, Combine or Merge
| Method | Behaviour |
|---|---|
| Replace | Replaces the table's data on each run |
| Combine | Adds new rows to the existing table without changing existing rows |
| Merge | Updates rows that match the Reference Column values and inserts rows with no match |
With Merge, choose at least one Reference Column: a column whose values uniquely identify each row, like a primary key. Where you set the method depends on the destination:
| Destination | Where you set it |
|---|---|
| Send to Dataset | In the form, under Each time this runs |
| Google BigQuery, PostgreSQL, MySQL, SQL Server, Amazon Redshift, Snowflake and Databricks (external) | Replace, Combine or Merge as the Update method |
| Google Sheets | Replace or Append, as the form's Write mode |
Send only new rows on each run
The Send only new rows checkbox, below When it runs, makes each run send only rows added since the last run.
- Tick Send only new rows.
- Under Using the column, pick a date or number column whose values only grow, such as an ID or a created date. Date columns are listed first, and the first one is preselected.
- Click Apply. The next run sends every row; later runs send only rows with a larger value than any already sent.
After a run, the form says which value the next run starts after. Click Start over and send every row next time to reset it. Picking a different column also starts over.
| Destination | Can send only new rows |
|---|---|
| PostgreSQL, MySQL, SQL Server, Amazon Redshift, Google BigQuery, Snowflake, Databricks | With Combine or Merge. With Replace: "Choose Combine or Merge to send only new rows. Replace would leave just the newest rows in the table." |
| Google Sheets | With Append only |
| Email, REST API Export | Yes |
| SFTP, SharePoint, OneDrive, Azure Blob Storage, Elasticsearch, Tableau Server/Cloud | No: "This destination is replaced on every run, so it can't send only new rows." |
If the View has no date or number column, the form says "Sending only new rows needs a number or date column whose values only grow, like an ID or a created date."
Related
- See when an export runs: triggers, Send at most, run history and failure emails.
- Hold exports for review: Auto-sync, the Monitor panel and troubleshooting.
- Send data to a database or warehouse: database and warehouse forms.
- Export to files and apps: file stores, Google Sheets and Tableau.
Something wrong? support@mammoth.io