Skip to content
Learn › Organise your data: open the docs menu

Find your way around the Mammoth Data Library

The Data Library holds a project's datasets, folders, connections, automations and Parameters. Click Data in the left sidebar to open it; a tab bar switches between Datasets, Connections, Automations and Parameters. Anyone with access to the project can browse it, and people with edit access can change what is in it.

This page covers the Datasets tab. Folders have their own guide: Folders and folder rules. Parameters and Snippets live on the Parameters tab: see Parameters and snippets.

What the rows in the Datasets list are

Click Data in the left sidebar. The Datasets tab opens on the top level of the project. Each row is one of four kinds of item:

RowWhat it is
FolderA container for datasets and other folders. The line under its name says how many items it holds, or "Empty folder".
DatasetA table of data. If it has more than one View, the line under its name gives the View count.
ViewOne View of a dataset. Views appear under their dataset when you expand it (see Change what the list shows).
Raw fileA file that is not a dataset yet, for example one that still needs a choice from you before it can be read.

Columns in the Datasets list

ColumnWhat it shows
NameThe item's name. A New tag marks an item that just arrived. An Interpreted tag marks a dataset read by the interpreter. A watched folder (one with automations that route files) says Watched.
StatusA chip that says what needs doing or what is happening (see the next section). Empty when all is well.
LinksHow the dataset connects to others: "fed by …", "joins …", "joined into …", "sends to …" or "exports to …". Click the link to see the names, then See them in the workflow to open the workflow map.
Cols | RowsThe number of columns and rows.
SizeThe size of the data.
UpdatedWhen the item last changed.

The list loads more rows as you scroll down.

What the status chips mean

The chip in the Status column tells you the state of a dataset and of the Views built on it. Chips marked "Yes" count towards the need-attention filter.

ChipWhat it meansNeeds attention
Processing or a count such as "2K of 10K rows"Mammoth is reading the data. A bar shows how far it has got.No
Adding data…New data is on its way into this dataset.No
Data not added or "3 files not added"A file added to this dataset did not fit its columns, so it is held until you review it.Yes
Reading…A PDF is being read the same way as one before it.No
Ready to readA PDF like one you already read is waiting for one click to read it the same way.Yes
Couldn't be readMammoth could not read the file.No
Needs attentionThe dataset needs a decision from you before it can be used. Click it to see what.Yes
"2 failing"That many Views of this dataset had a task fail on the last run.Yes
"2 out of date" or "3 batches behind"That many Views have not run since new data landed.Yes
New data waitingData pushed to the dataset's webhook is waiting to be pulled in.Yes
"1 sync off"That many Views have Auto-sync turned off, so they stay frozen when the dataset refreshes.No

When you expand a dataset, each View has its own chip: Failing, Out of date, Sync off or Updating. To fix these, see Control when Views update with Auto-sync.

Sort the Datasets list

Click a column header to sort by it. Click the same header again to reverse the order. The list opens sorted by Updated, newest first.

HeaderFirst click sorts
NameA to Z
StatusMost urgent first (folders last)
Cols | RowsMost rows first
SizeLargest first
UpdatedNewest first

Items with no number (a folder, or a dataset still processing) always sort to the bottom. Sorting applies to the rows loaded so far.

Show only the datasets that need attention

When at least one dataset needs attention, the List settings menu (the sliders button at the right of the toolbar) has a Filter section with an item such as "3 need attention". Tick it to show only those datasets.

The count is of datasets, not Views: a dataset with three failing Views counts once. The filter turns itself off when nothing needs attention any more.

Find a dataset by name

The Datasets tab has no search box of its own. Use Search in the top bar. It searches the whole workspace, and choosing a result opens that item in the Data Library.

Change what the list shows (List settings)

Click the List settings button (sliders icon) at the right of the toolbar.

SectionOptionWhat it does
RowsShow viewsExpands every dataset that has more than one View, listing its Views underneath. The hint under it says how many datasets have more than one View. It is greyed out when "Every dataset here has one view".
Filter"3 need attention"Shows only the datasets that need attention. Appears only when there are some.
DensityCompact, Comfortable, SpaciousRow height. Default: Comfortable.
ColumnsStatus, Links, Columns | Rows, Size, UpdatedTicks which columns show. All are on by default. Name always shows.

To collapse one expanded dataset, click the chevron beside it.

Your sort, density, columns and Show views choice are saved with your user preferences. The list looks the same the next time you open it, in any browser.

Select datasets, Views and folders

ToDo this
Open an item's panelClick the row. A dataset or View opens the dataset panel; a folder opens its folder panel.
Open an itemDouble-click it. A folder opens; a dataset or View opens in the data editor.
Tick one rowClick its checkbox, or Ctrl-click (Cmd-click on Mac) the row.
Tick a rangeTick one row, then Shift-click another. The range is added to what is already ticked.
Draw a boxPress on blank space and drag across rows. The box replaces the current selection with the rows it covers.
Tick everythingClick the checkbox in the header (Select all). Click it again to Clear selection.

From the keyboard:

  • the arrow keys move between rows;
  • Shift plus an arrow extends the selection;
  • Space ticks the focused row;
  • Enter opens its panel;
  • Shift+F10 (or the Menu key) opens its menu.

Act on several datasets at once

Tick two or more rows. The side panel becomes a selection panel titled, for example, "3 items selected". It lists the selection grouped under Datasets, Views, Folders and Raw files. Click the cross on a line to Remove from selection, or close the panel to clear it.

ActionWhat it doesWhen it is unavailable
CombinePuts the rows of the ticked datasets together. See Combine datasets."Only datasets can be combined." It is also blocked while a dataset is not ready.
Move toOpens the Move to… dialog to move everything into one folder. See Copy, move and delete datasets.The selection holds a View. A View moves only with its dataset.
DownloadDownloads the ticked datasets as one file (see below)."Raw files have no rows to download yet."
Move to trashMoves the items to Trash. For a folder, its contents go to Trash and the emptied folder is removed.The label becomes Delete when the selection holds items that cannot go to Trash, with the warning "2 items cannot go to Trash and would be deleted for good."

Every action waits while an item is still processing, for example "“Sales” is still processing."

Download several datasets as one file

Download opens a dialog where you choose the format:

OptionWhat you get
Excel workbookOne file with one sheet per dataset.
CSV files (.zip)A .zip file with one .csv file per dataset.
Stack the ones with the same columns into one fileDatasets with identical columns are stacked into a single file. The dialog says how many files you get.

Raw files are left out because they have no rows yet. When the file is ready it saves on its own. Limits are listed under Limits.

Every action in the More actions menu

Right-click any row, or click its ⋯ button, to see every action for that item. The menu has three bands: the main actions, then the other actions, then Move to trash in red.

ItemMain actionsOther actions
FolderOpen folderMove datasets in, Automate collection, Rename, Download folder, Move to
DatasetExplore, Export and publish, the automation actionAdd data (or Get the latest / Retrieve for a linked or connector dataset), Edit settings (connector datasets), New view, Rename, Settings, Clone dataset, Move to, Download
ViewExplore, Export and publish, the automation actionRetrieve (connector datasets), Rename, Download
Raw filenoneMove to

A raw file cannot go to Trash, so its last item is Delete permanently. A dataset that needs a decision shows Resolve first. While an item is processing, the menu offers nothing that changes it until processing finishes.

The dataset actions are described in The dataset panel. The folder actions are in Folders and folder rules.

Add a new dataset or folder

The toolbar above the list has two buttons:

ButtonWhat it does
New datasetOpens the add-data dialog, with every way in: file upload, connectors, URL, paste, sample data, Generate with AI and webhooks. See Choose how to bring data in.
New folderCreates a folder in the folder you are in. See Folders and folder rules.

New datasets go into the folder you are in. While a dataset is being created, a placeholder row with a spinner sits at the top of the list. On a computer, the new dataset's panel opens when it arrives.

Drop files onto the list

Drag files from your computer anywhere onto the list. The upload dialog opens and takes the files, ready to upload into the current folder. See Upload files.

Start from an empty library

A project with no datasets shows "No Datasets yet" with an upload area. Drop files on it or choose them, then click Upload. Three buttons offer the other ways in: Connect to a source, Use a sample and Generate with AI.

Use the Data Library on a tablet or phone

The Datasets tab adapts to narrow screens:

Screen widthLayout
960 px and widerThe list and the panel sit side by side.
600 to 959 pxThe panel slides over the list as a drawer. Tap outside it to close it.
Under 600 pxThe panel replaces the list. Each row shows one line of columns, rows and when it was updated. A sort button above the list (for example "Sort: Updated") picks the sort column and order.

On a screen under 600 px wide:

  • long-press a row (or choose Select in its menu) to start selecting;
  • while anything is ticked, a tap ticks or unticks;
  • the + button opens New dataset and New Folder.

On a tablet or phone, a new dataset doesn't open by itself. A notice says, for example, "“Sales” just arrived." or "“Sales” is ready." with Open and Dismiss.

Who can do what in the Data Library

Anyone with access to the project can browse the list, sort and filter it, open panels and explore data. With view-only access, actions that change data are greyed out: "You have view-only access to this workspace, so you cannot add data." The selection panel says "You have view-only access to this workspace, so these actions are unavailable."

For project roles, see Members, roles and custom roles.

Limits

LimitValue
Datasets in one download200 ("One download holds at most 200 datasets. This one has 240.")
Size of one download1 GB ("This download is over 1 GB. Pick fewer datasets.")

Troubleshooting

SymptomCauseFix
"This list could not be loaded. Your data has not changed."The list failed to load.Click Try again.
Combine says "Only datasets can be combined."The selection holds a folder, View or raw file.Untick everything that is not a dataset.
Move to is unavailableThe selection holds a View.Untick the View, or move its dataset.
An action waits with "“Sales” is still processing."An item in the selection is still processing.Wait until processing finishes.
Was this page helpful?

Something wrong? support@mammoth.io