Table and JSON views

The toggle above the results offers Table, JSON and 📊 Chart. The choice is remembered per tab. Table is best for scanning many items and editing scalar values; JSON mirrors the MongoDB Compass experience for nested documents; Chart turns the loaded page into timelines, histograms and a partition heatmap (see Charts and heatmap).

Columns are discovered from the loaded items: the partition key comes first, the sort key second and the remaining attributes follow alphabetically. When later pages reveal new attributes, their columns are added. Key columns are styled distinctly and cannot be edited inline.

• Sort — click a header to cycle ascending, descending, off. Sorting is client-side on the loaded page: numbers compare numerically, strings compare case-insensitively with numeric awareness, and rows lacking the attribute sort last.
• Reorder — drag a header by its ⠿ handle. The custom order lasts until the next Execute or Refresh, which re-sorts columns.
• Resize — drag the right edge of a header. Widths are remembered by column name for the session.
• Cells — null is shown in italics, objects and arrays as compact JSON, booleans and numbers colour-coded. Long values are clipped at 300 px; hover for the full value.

Every row ends with an actions column:

Each row has a checkbox, and the header checkbox selects every visible row (it shows an indeterminate state for partial selections). Hold ⇧ and click to select a range from the last clicked row. Selection is tracked by primary key and limited to the current page.

With one or more rows selected, the toolbar shows a counter, a 🗑 Delete button for bulk delete and a ✕ to clear the selection. Selection is cleared automatically when you switch tables or after a bulk delete.

Every item is rendered as a syntax-highlighted, collapsible tree. Click a { or [ line to fold that object or array; arrays show an item-count hint. Each record has its own actions in the top-right corner: ✎ Edit Item, 📄 Clone Item, copy JSON and 🗑 Delete Item.

A format toggle at the top switches between JSON (native values) and DynamoDB JSON (typed attribute values). The setting is global and remembered. Copy follows the displayed format; Edit, Clone and Delete always operate on the native item.

{
  "userId": "usr_001",
  "email": "alice@example.com",
  "active": true,
  "tags": ["admin", "beta"],
  "profile": { "plan": "pro", "seats": 3 }
}

The toolbar search box filters the items already loaded on the current page by matching your text against the entire serialised item, including nested values and attribute names. It is instant, works in both Table and JSON views, is remembered per tab and never calls DynamoDB. Focus it with ⌘+F.