Datasets

Find the data you've imported, see who can use it, share it with your organization, and organize the tables inside each dataset.

Where: Datasets

A dataset is data you've brought into eode from an Excel workbook. You import it once and it stays in the Dataset Library, where you open it, edit it, and pick it for any model. A dataset's tables are called resources.

Find a dataset

Click Datasets in the rail, just below Model Library. The library works with no model open.

The strip at the top left picks which datasets are listed. Each tab carries a count.

Tab	Lists
PERSONAL	Your own datasets, plus personal ones whose roster names you. The library opens here on every visit.
ORGANIZATIONAL	Datasets promoted to your whole organization.
ALL	Everything you can see.

Search datasets… matches the dataset name only, not resource names. Switching tabs clears it.
Sort orders the list by Last updated (the default), Date created or Name. Dates sort newest first. The choice resets when you leave.
The tab counts don't shrink while you search.

Each row shows the name, a PERSONAL or ORGANIZATIONAL badge, and a line like 2 resources · 6 rows. Click the name to open the dataset.

The Harbor Point Bakery — sales and costs row: its PERSONAL badge, 2 resources · 6 rows, and the rename, duplicate, permissions, promote and delete buttons

Act on a dataset from its row

The buttons on the right of every row are always visible. In a narrow window they drop below the name, still on the right, so a long name is never squeezed out.

Button	What it does
Pencil	Rename. An empty or unchanged name does nothing.
Copy	Duplicate. "Its resources and rows are copied into a new dataset you own. Edits to the copy do not touch the original." The name defaults to `<name> (copy)`. The copy opens when it's made.
Shield	Permissions. Shows who holds which role on the dataset. See below.
Globe	Promote to Organizational or Demote to Personal. Only for some roles; see below.
Trash	Delete. See below.

If a duplicate, promote or demote fails, a line above the list says Could not duplicate —, Could not promote — or Could not demote —, followed by the reason.

See who can use a dataset

Click the shield on a row. Permissions · name lists the people in each role: Owners, Contributors, Viewers, Promoters and AI agents, each with a count. An empty role reads — none —.

The view is read-only. "This view is read-only — it mirrors the recorded permission map for review." + Add is turned off.

What each role can do:

You are	You can
Anyone who can see the dataset	Open it, preview it, duplicate it, pick it for a model.
Owner, contributor, or Org Admin	Also rename it, edit its resources and save.
Owner or Org Admin	Also delete it.

You see a dataset when its roster names you, or when it's organizational and belongs to your organization. An Org Admin sees every dataset in the organization. Changes to an organizational dataset you aren't on the roster of are refused when you save.

Org Admin or Promoter

Click the globe on a personal dataset, then Promote. "Every member of your organization will be able to see and preview it. Only its roster can edit it." It moves to the ORGANIZATIONAL tab.

On an organizational dataset the globe reads Demote to Personal. "It goes back to your personal library; only the people on its roster keep access."

The globe shows only to Org Admins and Promoters. Promoting also needs you to be the dataset's owner, a promoter on it, or an Org Admin. Roles are explained in Roles and personas.

Delete a dataset

Click the trash on a row. Before it asks, eode counts the models built from the dataset. The dialog Delete “name”? starts with that count when there are any: "N models were built from this data. A model that loads it at Load data will need its workbook loaded again." Then: "Its imported resources are removed with it. This cannot be undone."

A model that reads its workbook from the deleted dataset asks for the workbook again the next time it loads data; see Load your data.

Open a dataset

Click a dataset's name. The page shows the name, the resource count, and two columns: the Resources list on the left and the open resource's grid on the right.

The Harbor Point Bakery dataset page: 2 resources, the Resources list with Products open, and the Products grid with its toolbar and footer

The first resource opens by default. The one you last picked stays open after a save or a reload.
Each resource in the list shows its row count. Click one to open it.
Editing the grid is covered in Edit a dataset.

Switching to another resource drops unsaved edits in the grid without asking. Save first.

When a resource was cut at the row limit during import, a line above its grid reads "Truncated from N rows at the source.", with N the row count in the workbook. See Import limits.

Organize the resources

The top of the Resources list has three controls.

Control	What it does
Filter resources	Narrows the list by name as you type. `Esc` clears it, a second `Esc` leaves the box, `↓` moves into the list. The open resource stays open even when it's filtered out.
Sort resources	Name A–Z (the default) or Name Z–A. Your pick is remembered in this browser.
+ (Add resource)	Empty table or Import data.

Hover a resource to show its pencil (Rename resource) and trash (Delete resource). Deleting asks first: "Its rows are removed with it. This cannot be undone."

The Add resource menu open over the Resources list, offering Empty table and Import data

Add an empty table

Click +, then Empty table.
Type a name in New resource ("It starts empty, with a header row you can rename in the grid.") and click Next.
Enter how many columns it should have (3 by default) and click Create.

The table opens with columns named Column 1, Column 2, and so on, and one blank row. Rename the columns in the grid.

Add tables from a workbook

Click +, then Import data. The Import Data window opens without a name field, and the tables you pick are added to this dataset.

When a list is empty or doesn't load

You see	What it means
No personal datasets yet.	Nothing in the PERSONAL tab. Import Data starts one.
No datasets yet.	Nothing in the ALL tab: you have no datasets and none are shared with you. Import Data starts one.
No organizational datasets yet.	Nobody has promoted a dataset in your organization.
No datasets match your search.	Clear search shows the whole tab again.
Could not load your datasets — reason	The list failed to load. Try again.
This dataset has no resources.	The import produced no data. New resource adds an empty table.
Could not load this dataset — reason	Try again.
Could not create the resource —, Could not rename the resource —, Could not delete the resource —	The action was refused or failed; the reason follows.

Limits

Search matches dataset names only.
An empty table starts with at most 50 columns. A larger number is cut to 50. Add more in the grid.

Import data: bring a workbook in as a new dataset, or add its tables to one
Edit a dataset: change rows, columns and values in a resource
Load your data: pick a dataset for a model and bind it
Data connections: the Model Library tab for connected sources
Model Library: the same tiers for models

On this page