Folders
Folders group the datasets in a project's Data Library. A folder can hold datasets, raw files and other folders, to any depth. This guide is for anyone who organizes a project's data. For the list itself, see Data Library.
Create a folder
- Click Data in the left sidebar. The Datasets tab opens.
- Open the folder you want the new folder to sit in, or stay at the top level.
- Click New folder in the toolbar. On a phone, tap +, then New Folder.
- In the New folder dialog, type a name under Folder name.
- Click Create folder, or press Enter.
The folder appears in the list and a message such as "Created new folder 'Sales'" confirms it.
| Rule | Detail |
|---|---|
| Name length | Up to 100 characters. |
| Unique names | Two folders in the same place can't share a name (the check ignores case). The dialog says, for example, "A folder named “Sales” already exists here" and Create folder stays greyed out. The same name is fine in a different folder. |
| Empty name | Create folder stays greyed out until you type a name. |
| Nesting | Create a folder inside another by opening the parent first. |
You can also create a folder while moving items, from the Move to… dialog. See Cloning and moving datasets.
Open a folder and find your way back
- Double-click a folder row, or click it and choose Open folder, to step inside.
- Inside a folder, the breadcrumb above the toolbar (Folder path) shows where you are. It starts with Datasets (the top level of the project). Click any crumb to jump there, or click the back arrow (Back to the folder above).
- Each folder has its own address, so you can bookmark it, reload it, or share the link with someone who has access to the project.
An empty folder shows an upload area. Drop files on it, or use Connect to a source, Use a sample or Generate with AI. While a folder is still being filled (for example after you upload a folder of files), it says "Preparing N files" and the datasets appear as they finish.
The folder panel
Click a folder row once to open its panel beside the list. Under Contents a table lists what the folder holds, with Name, Type, Rows and Updated. Click a row in that table to open the item: a folder opens, a dataset opens in the data editor.
The panel and the folder's ⋯ menu (or right-click) offer these actions:
| Action | What it does |
|---|---|
| Open folder | Steps into the folder. |
| Move datasets in | Moves the datasets you ticked in the list into this folder. |
| Automate collection | Sets up an automation that drops new files into this folder on a schedule, from SFTP or a connection. |
| Rename | Lets you edit the folder's name in the panel header. |
| Download folder | Downloads every dataset inside as one file. |
| Move to | Moves the folder, with everything in it, to another folder. |
| Move to trash | Moves the folder's contents to Trash. The empty folder is then removed. |
With view-only access you can open folders but not change them.
Rename a folder
- Right-click the folder (or click its ⋯ button) and choose Rename. The folder's panel opens with its name ready to edit.
- Type the new name and press Enter.
The same rules apply as when you create a folder: up to 100 characters, and no other folder in the same place may have the same name. If it does, you see a message such as "A folder named “Sales” already exists here" and the old name stays.
Move datasets into a folder
There are three ways:
| Method | How |
|---|---|
| Move datasets in | Tick the datasets in the list. Then open the target folder's ⋯ menu and choose Move datasets in. If nothing is ticked, Mammoth says "Tick the datasets in the list first, then choose Move datasets in." |
| Drag and drop | Drag one row, or a ticked set of rows, onto a folder row or a crumb in the breadcrumb. A message such as "3 items moved to Sales" appears with Undo. |
| Move to | Choose Move to on the item (or on a selection) and pick the folder in the Move to… dialog. |
A view can't be moved on its own: it moves with its dataset. A folder can't move into itself or one of its own subfolders. For the Move to… dialog in detail, see Cloning and moving datasets.
Download a folder
-
Right-click the folder and choose Download folder.
-
Pick a format:
Option What you get Excel workbook One sheet per dataset. CSV files (.zip) One .csv file per dataset, in a .zip file. Stack the ones with the same columns into one file Datasets with identical columns are stacked into one file. -
Click Download. The file saves on its own when it is ready.
One download holds at most 200 datasets and 1 GB. Raw files and tables that can't be read yet are left out, and the dialog says how many.
Delete a folder
- Right-click the folder and choose Move to trash.
- The Move to trash? dialog says what goes with it: "The N items inside leave the library now and are recoverable from Trash for 30 days." (The number of days is your workspace's Trash retention.)
- Confirm.
The folder itself isn't a Trash item: Trash lists the datasets and Views that were inside it, and the emptied folder is removed. To get them back, restore them from Trash, and each dataset returns to its original folder. See Trash and recovery.
Folders with automations
A folder can be the source of a File Collection or Data Consolidation automation. Such a folder is Watched: files that arrive in it are routed into datasets by rules. Set these automations up from Automate collection or the Automations tab. See Automation types and Data Consolidation.
What the folder tells you
Inside a watched folder, the toolbar shows a box for each of its automations, after New folder. Click a box to see:
| Field | What it shows |
|---|---|
| Takes | Which files the rule takes: all files, or names containing, starting with, ending with, or exactly matching some text. |
| Goes into | The dataset (or datasets) the files go into, and where that dataset lives. |
| Last file / Last run | When the automation last took a file or ran. |
| Active / Paused | Whether the automation is running. |
Click View automation to open the automation. Its History lists every file it took in, how it arrived (Uploaded file, Collected file, Moved in, Added on request) and how it went (for example "2 of 3 added").
The card also says: "Once a file is added it leaves this folder. A file that can't be added stays here, and we'll tell you why."
In the list, a watched folder's row says Watched and which dataset it feeds. The destination dataset's row says "Fed by" and the folder's name.
Change the order of a folder's rules
When a file matches more than one rule, it goes by the first rule it matches ("Files added here go by the first rule they match"). Anything no rule takes stays in the folder as its own dataset.
- Click the box that counts the folder's automations (for example "2 automations").
- Click Reorder rules. This appears only when the folder has two or more rules of its own.
- Use Move up and Move down to set the order.
- Click Save order. A dialog asks "Change the order of this folder's rules?" and explains that only new files follow the new order. Click Save order again.
The message "Rule order saved" confirms it. If it fails: "Couldn't save the order. Nothing changed." Rules set on a parent folder show "from" and that folder's name, and can't be reordered here.
Where these files go
When you upload files into a watched folder, the upload dialog shows Where these files go before you upload. Each file is grouped under the dataset it will go into (for example "→ Sales"), or under Stays here ("No rule takes it, so it becomes its own dataset here."). If a file matches two rules, it says which comes first.
When a file doesn't go in
A file that couldn't be added stays in the folder with a chip and a reason. Some chips come with a button that fixes it.
| Chip | What it means | Button |
|---|---|---|
| Reading… / Adding… | The file is on its way into its dataset. | none |
| Not added | The file didn't fit, for example "Its columns don't match Sales" or "Has a new column for Sales: Region". | Match the columns for column problems, Review otherwise |
| Pick sheets | A workbook needs you to choose which sheets go in. | Pick sheets |
| Duplicate | The file has the same contents as one added earlier, so it was skipped. | Add anyway |
| Needs password | The file is protected and waits for its password. | Review |
| Kept separate | The dataset's recipe couldn't read the file, so it became its own dataset. | Review |
| Didn't go in | Adding failed, for example "It stopped on its way to Sales; add it again". | none |
Other reasons you may see include "Not read: this month's AI pages for PDFs are used up", "Not read: PDFs need an AI key, and none is set up" and "This kind of file can't be read". The folder row counts the files waiting ("N not added").
Related articles
- Data Library: the Datasets list, selection and bulk actions
- Cloning and moving datasets: the Move to… dialog in detail
- The dataset panel: what you can do with a dataset
- Trash and recovery: restore the contents of a deleted folder