> 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.

# Add rows to a dataset or replace its data in Mammoth

> Use Add data to append rows to a Mammoth dataset or replace its data from a new file or another dataset. Covers update method, column mapping and validation.

The **Add data** dialog loads new data into a dataset you already have. You can add new rows on top of the existing data, or replace it, from a file upload or from another dataset in the project. Open the dataset's panel in the **Data Library** and click **Add data** ("Append a batch, or swap the file for a newer one"). You need Edit-level access to the dataset.

Use it to:

- load the latest period into a running dataset (add);
- swap a dataset's data for a refreshed export (replace);
- merge two datasets that share the same columns.

To put the rows of two or more datasets together from the list (tick them, then **Combine**), see [Combine datasets](https://mammoth.io/docs/learn/organise/combine-datasets/). Its **Batches in one dataset** option finishes in this dialog.

> Note: The operation runs in the background. You don't need to wait on the dialog. Track progress in the **Batches** section of the dataset panel (under **Settings & history**) or in the **Activity Log**.

## Open the Add data dialog

1. In the Data Library, click a dataset row to open the dataset panel.
2. Click **Add data**.

You can also choose **Add data** from the row's **⋯** menu, or from the **Batches** section under **Settings & history**.

**Add data** is offered for datasets you upload files into. Datasets fed by a connection offer **Retrieve** instead, and datasets that follow a file link offer **Get the latest**.

The dialog is titled **Add data**, with the dataset name beneath. It has an **Input file** row and an **Update method** row.

![Add data dialog for Daily Sales.csv with Input file (Upload file or Select existing), Update method Add to existing data, the drop zone and Column mapping](https://mammoth.io/docs/images/20261102_add-data-dialog.webp)

| Your situation | **Input file** | Go to |
| --- | --- | --- |
| You have a new file (CSV, Excel and so on) to load | **Upload file** | [Add a file to a dataset](#add-a-file-to-a-dataset) |
| You want data from another dataset in your project | **Select existing** | [Add another dataset's rows to a dataset](#add-another-datasets-rows-to-a-dataset) |

## Choose add or replace: the update method

| Method | What it does | Best for |
| --- | --- | --- |
| **Add to existing data** | Appends new rows to the existing dataset. Previous data is kept. | Incremental loads such as daily sales, event logs, transaction feeds |
| **Replace all data** | Deletes all existing rows and inserts the new data | Full snapshots such as weekly inventory exports or complete dimension tables |

Changing the **Update method** saves it to the dataset straight away, even if you then select **Cancel**: "This changes the update method of the whole dataset, not only this upload. It stays changed even if you cancel adding the file." The same setting applies to both **Upload file** and **Select existing**.

![Update Method dialog: Replace all data or Add to existing data, with the warning that the change applies to the whole dataset](https://mammoth.io/docs/images/20261102_update-method-options.webp)

Previous data can't be recovered once replaced. Consider keeping a copy of important history before you replace it. See [Batches: see and fix what arrived](https://mammoth.io/docs/learn/get-data-in/batches/#combine-or-replace-how-new-data-joins-old-data).

## Add a file to a dataset

Prerequisite: Edit-level access to the dataset, and a file in a supported format.

1. Drag a file onto the drop zone, or click **Upload file** to browse your computer.
2. Choose the **Update method**: **Add to existing data** or **Replace all data**.
3. Optionally set **Column mapping** (see the next section). This row is not shown for PDF files.
4. If the update method is **Replace all data** and the file is not a PDF, select **Replace and use the new file's columns** when the file has new or renamed columns. The dataset's columns then become the file's columns.
5. Click **Submit**. The button stays disabled until a file is selected.

A callout in the dialog explains that uploading a PDF replays the dataset's saved recipe automatically, while uploading a CSV or Excel file appends it directly by matching columns.

| Detail | Value |
| --- | --- |
| Maximum file size | 50 MB ([Limits and quotas](https://mammoth.io/docs/learn/help/limits/)) |
| Accepted formats | The same as for [file upload](https://mammoth.io/docs/learn/get-data-in/upload-files/#supported-file-formats-for-upload): CSV, TSV, PSV, TXT, XLS, XLSX, XLSM, XML, JSON, JSONL, NDJSON, ZIP, BZ2, GZ, TAR, 7Z, PDF, TIFF, TIF, JPEG, JPG, PNG, HEIC, HEIF, WEBP |

Result: the dialog closes and Mammoth processes the file in the background. A new row appears in the **Batches** list. It reads **Applying recipe** while Mammoth structures the file, or **Pending** if the batch needs your input, such as a column mapping, a sheet selection or a file password. Select a pending row to resolve it (see [Resolve a pending batch](https://mammoth.io/docs/learn/get-data-in/batches/#resolve-a-pending-batch)).

## Set column mapping and type rules for new files

The **Column mapping** line appears below the file drop zone. It shows whether the dataset's incoming-data rules are **Default** or **Custom**:

- **Default:** Mammoth maps columns by matching names and converts mismatched column types to text.
- **Custom:** the dataset has saved, non-default mapping rules.

![Column mapping row showing Default, with a pencil icon at the right to edit](https://mammoth.io/docs/images/20260505_060023_batch_col_mapping_area1.webp)

Select the row (a pencil icon appears when you hover) to open the **Combine / Replace Settings** dialog.

![Combine / Replace Settings dialog with If column type doesn't match set to Convert destination column to text and Column Mapping set to Map with existing columns](https://mammoth.io/docs/images/20260505_060336_combine_replace_setting_modal1.webp)

### If column type doesn't match

| Option | What it does |
| --- | --- |
| **Convert destination column to text** (default) | Changes the destination column's type to Text so all incoming values fit. Broadest compatibility, with no data lost. |
| **Convert incompatible values to null** | Keeps the destination column's type. Incoming values that can't be cast to it are stored as `null`. |
| **Do not update the dataset** | Stops the operation if any column type mismatch is found. Use it when strict type integrity matters. |

### Match columns with different names using aliases

| Option | What it does |
| --- | --- |
| **Map with existing columns** (default) | Matches incoming columns to destination columns by exact display name. No manual mapping needed. |
| **Create new column mapping** | You define one or more name aliases per destination column. Mammoth uses them to find the right incoming column even when the name differs between files. |

Use aliases when the same column has different headers in different files, for example `Cust_ID`, `CustomerID` or `Customer id`. With **Create new column mapping** selected, a table lists every destination column with its aliases:

| Element | What it does |
| --- | --- |
| Alias pills (for example, `Customer id ×`) | The names Mammoth looks for in incoming data. Click a pill to edit its name; click **×** to remove it. Several aliases are joined with "or", so any match works. |
| **+** button | Adds a new alias for this column |
| Toggle | Turns mapping on or off for this destination column |
| Type icon and column name (right side) | The destination column's data type and name |

Tick **All destination columns must be mapped** to add new data only when its number of columns matches the existing dataset.

Click **Save changes**. The **Column mapping** line now shows **Custom**, and future loads use your saved rules.

## Add another dataset's rows to a dataset

Prerequisite: Edit-level access to the destination dataset, and the source dataset is in the same project.

### Pick the source dataset

1. Choose **Select existing**.
2. Open the **Dataset** picker and choose the source. It lists the other datasets in the project, plus the project's unprocessed PDF files (marked **Unprocessed PDF**).
3. Check the data preview of the source columns and first rows.

![Select existing source list showing store_transaction and cust1-extra-col.xls_test02, a data preview, and a Review Column Mapping button](https://mammoth.io/docs/images/20260430_065845_select-existing-list-dataset-to-combine1.webp)

Datasets that hold unstructured data are marked with a warning icon and **Needs structuring**. They appear only when the destination dataset has a saved recipe, and they must be structured to match the destination before they can combine.

Two picks behave differently:

- **Unprocessed PDF:** Mammoth replays the destination's saved recipe on the PDF. There is no column mapping, and the button is **Submit**.
- **Needs structuring:** the button is **Apply Recipe**. It previews the destination's recipe applied to the source dataset, and nothing is written until you select **Submit**. Submitting confirms the recipe on the source and deletes the source dataset.

### Check the update method

The **Update method** at the top of the dialog applies here too. **Add to existing data** appends the source rows and keeps previous data. **Replace all data** deletes all existing rows and inserts the source rows.

![Animation of the Add or Replace data dialog with Select existing chosen and an empty Select dataset dropdown](https://mammoth.io/docs/images/20260508_050328_map-replace.webp)

### Keep or delete the source dataset

The **Delete dataset after combining** checkbox is ticked by default. It is hidden when you pick a PDF or a dataset that needs structuring. If left ticked, Mammoth permanently deletes the source dataset after the operation. Untick it to keep the source. Leave it ticked only when the source is a temporary staging dataset you no longer need.

![Add or Replace data dialog with Select existing chosen, orders.csv in the source field, and Delete dataset after combining checked](https://mammoth.io/docs/images/20260505_060220_Screenshot2026-04-281612071.webp)

### Check that the columns match

After you pick a source, Mammoth checks whether its columns match the destination's.

| Outcome | What you see | What to do |
| --- | --- | --- |
| Columns match | Green check: "New data matches the settings" | Click **Validate** or **Submit**, or **Review Column Mapping** to adjust |
| Columns differ | Red message: "These columns do not all match the dataset's, by name or by type. Review the column mapping to line them up." | Click **Review Column Mapping** and map the columns |

### Map the columns when they differ

![Column mapping view with Customer id unmapped, Customer name and Item matched, Save results into options, and Previous and Validate buttons](https://mammoth.io/docs/images/20260430_071019_review-column-mapping1.webp)

The sub-header confirms what is happening, for example "Daily Sales is combining data with orders.csv", with the **Delete dataset after combining** checkbox beside it.

The mapping table lists every destination column on the left (with its data type icon: `T` for Text, `#` for Number, `⏱` for Date / Time) and a **Select column** dropdown on the right for the matching source column.

- Use **Filter columns by** to show **All**, **Matched** or **Unmatched** columns.
- A column search appears when the dataset has many columns or when you start typing.
- When a mapped source column has a different type from the destination column, a selector lets you choose which type to keep.
- **Validate** is enabled once at least one column is mapped (and, for a new dataset, once it has a name). Unmapped columns are allowed.
- After a successful validation, the button becomes **Submit**.

### Save the result as a new dataset

By default, the combined result overwrites the destination dataset. Set **Save results into** to **New Dataset** to keep both the source and the destination and write the output to a new dataset. You can name it (up to 300 characters). With **New Dataset**, **Submit** is available without validating.

## Check what the change affects before you submit

When you click **Validate**, Mammoth checks for Views, pipelines or tasks that depend on the destination dataset and would be affected. If it finds any, a **Review** window lists them. Confirm you understand the impact before the final submit becomes available.

## What happens after you submit

The dialog closes straight away and the operation runs in the background. For **Select existing**, a message confirms when the source dataset will be deleted. Mammoth:

1. Appends or replaces the rows in the destination dataset.
2. With **Replace all data**, deletes the previous batch records.
3. With **Delete dataset after combining** ticked, removes the source dataset.
4. Updates the batch count in the dataset panel.
5. Reruns dependent Views and pipelines. A View whose incoming link has Auto-sync off is not rerun: the new batch stays pending, and Monitor shows "new batch received, auto-sync off" with **Preview** and **Process** buttons (see [Views](https://mammoth.io/docs/learn/organise/views/)).
6. Writes an Activity Log entry recording what happened, who triggered it and when.

With **Add to existing data**, the batch count goes up by one per operation. With **Replace all data**, the previous batches are removed and one new batch appears. It is created suspended if all the previous batches were suspended.

Follow progress in the **Batches** section of the dataset panel or in the [Activity log](https://mammoth.io/docs/learn/automate/activity-log/).

## Troubleshooting adding and replacing data

| Symptom | Cause | Fix |
| --- | --- | --- |
| **Add data** isn't offered | The dataset is fed by a connection or follows a link | Use **Retrieve** or **Get the latest** instead |
| "These columns do not all match the dataset's, by name or by type." | Column names or types differ | Click **Review Column Mapping**, or set aliases with **Create new column mapping** |
| The new batch shows **Pending** | It needs a column mapping, a sheet selection or a file password | Select the row to resolve it (see [Resolve a pending batch](https://mammoth.io/docs/learn/get-data-in/batches/#resolve-a-pending-batch)) |
| The update method changed although you cancelled | Changing **Update method** saves straight away | Set it back in the dialog or in the dataset's settings |
| A View didn't update after the new data | Auto-sync is off on its incoming link | Click **Process** in Monitor, or turn Auto-sync on (see [Views](https://mammoth.io/docs/learn/organise/views/)) |
| The source dataset disappeared | **Delete dataset after combining** was ticked, which deletes the source permanently | Untick it next time you want to keep the source |

## Related

- [Batches: see and fix what arrived](https://mammoth.io/docs/learn/get-data-in/batches/): every load is a batch you can suspend or delete.
- [Combine datasets](https://mammoth.io/docs/learn/organise/combine-datasets/): put the rows of several datasets together from the list.
- [Combine incoming data](https://mammoth.io/docs/learn/automate/consolidate/): add or replace from new files on their own.
- [Refresh a dataset](https://mammoth.io/docs/learn/automate/refresh-a-dataset/): refresh a connected dataset on a schedule with the same add or replace logic.
- [Activity log](https://mammoth.io/docs/learn/automate/activity-log/): every batch operation with time and user.

---
Source: https://mammoth.io/docs/learn/get-data-in/add-or-replace-data.md · Updated: 2026-11-08