The Data tab shows the objects in your space by Tag: your pages, tasks, and everything you tagged. The header shows how many objects the tag has and whether it has a schema; click that to open its schema. For the selected tag it has two views: Objects, where a query editor narrows, sorts and groups the objects without you having to know the query syntax, and Schema, where you look at and edit the tag's attributes.

${widgets.commandButton("Workbench: Data")} opens this tab in the Workbench, and ${widgets.commandButton("Workbench: Schema")} opens it on Schema.

The tag list

• Pinned tags come first: page and task unless you change them. Hover a tag and click its pin to pin or unpin it. Pins are kept in the workbench.pinnedTags option of your ^Library/Config, so everyone using the space sees the same ones.
• Defined: your tags that have a Schema, most objects first. A tag with objects that don't match its schema shows a ⚠ count. A tag that's defined but not used yet is grayed out.
• No schema: your tags without one, most objects first, so the list reads as a to-do list.
• Built-in: the tags SilverBullet indexes itself (item, link, header, …), collapsed until you open it.

Tags from data blocks carry a small data label. Drag the list's right edge to make it wider; double-click the edge to reset it. Typing in the filter narrows every section; Enter picks the first match. Each tag shows its icon: a tag's icon, or else the icon of its pageDecoration (see API/tag).

Objects

The Objects view has three parts:

• The query editor: one row per part of the query, named after the keyword it writes.
• The results: rendered the way the query renders on a page (or as Lua values). It shows the first 200 rows of a large result.
• The query bar at the bottom.

Building a query

• From: the tag picked in the list. For page and task, tagged narrows them to one of your tags, and Open #tag jumps to that tag.
• Where: filters, all of which must match:
  • + Filter picks an attribute, then an operator that suits its type.
  • + Expression takes any Lua condition.
• Group By: one result per distinct value of an attribute (or combination of attributes). The Select row then shows group columns: each is a group key or an aggregate of an attribute, like count or avg, including your own aggregates.
• Order By: the order, with directions named for the attribute’s type (earliest first, A–Z, high to low). Click an order to flip it. While grouping, you order by a group key or an aggregate column.
• Select: what each result shows.
  • Table (the default): a column per attribute the objects have. Remove, rename and reorder columns, or Clear all to start from none. A Page link column makes an attribute a link, and an Expression column computes its value.
  • List: one line per result, from a template your space defines under templates.* (such as tasks with their checkbox), or a Custom template… you write with the attributes as ${name}.
• Limit: how many results you’d like to show.

Changing the tag, or a change that leaves a filter or column without its attribute, offers to undo it.

Using the query

The bar at the bottom shows the query (at least part of it).

• Insert at cursor puts the ${query[[…]]} in the page you’re editing, or where you typed /Query…. For a query you opened from a page, it’s Update page, which replaces that query, and keeps doing so as you change it further.
• Copy copies it.
• Run in Console runs it in the Console, where you can look at its result as Lua values.

Editing a query on a page

To change a ${query[[…]]} already on a page, choose Edit in query builder from its widget’s menu, or put the cursor in it and run ${widgets.commandButton("Workbench: Edit Query")}. This is best effort: not every handwritten query can be represented by the visual builder. Editing and inserting queries requires a writable page. For a complete exercise in your own space, follow Guide/Add Structure and Query.

Schema

The Schema view lists the tag's attributes: the ones its schema declares, and the ones found by looking at up to 500 of its objects. For each it shows:

• Attribute and Type: the JSON Schema type (string, number, integer, boolean, array, object, null), a format such as date, allowed values (enum), a reference to another tag (ref), or a list of tags or @names (tags, identities). allow empty also accepts the attribute left empty, such as due: with nothing after it (a null value). That's separate from Required, which says whether the attribute must be there at all. Rules the editor can't show as simple controls (an anyOf of different types, say) offer Edit as Lua, also in every row's ⋯ menu: the rule as a Lua table or schema.* helpers, read back without running it.
• Description and Required.
• Present: how many of the sampled objects have a value. A partial count opens the objects without it. A ⚠ invalid count opens the objects whose value doesn't match, and an allowed value opens the objects that have it.
• Origin: inherited from the root tag (a #book page has name because it's a page), declared in a definition elsewhere, local to your changes here, or inferred from the objects.

Edit any of them in place, add attributes with Add attribute, and remove or reset them from a row's ⋯ menu, then click Save and apply (or press Mod-s). Changes stay pending, also when you switch tags or close the Workbench, until you save or discard them; the tab shows a • meanwhile. Copy as tag.define copies the schema as Lua, and the icon button picks the tag's icon (or No icon). For a schema defined elsewhere, the toolbar links to the page that defines it.

For a tag without a schema, tick the inferred attributes to keep and click Create schema. Nothing is required until you say so.

Renaming or removing an attribute changes only the schema, never your pages: an attribute you removed shows up again as inferred while objects still have it.

Where schemas are saved

The Data tab saves to its own block in your ^Library/Config, marked -- managed-by: workbench-schemas, next to the block the Configuration tab keeps. It's ordinary Space Lua you can read and edit:

• A schema created here is a tag.define.
• Changes to a schema defined elsewhere, such as adding an attribute to page or to a library's tag, are kept as separate config.set and tag.setRequired calls. The original definition is left alone, so later changes to it still take effect.

The block runs after the built-in tags and your other scripts (-- priority: -10). When a script that runs even later replaces a schema, saving says which of your changes it overrides.