Plota docs
Product GitHub Open the editor

7. Data and conditional formatting

Left panel → Data (or the palette, Shape data: open the panel).

The data panel
The data panel

Shape data#

Any node can carry typed key/value properties. Types are string, number, boolean, date and url. Add, rename, retype and delete them in Data ▸ Shape data with one node selected.

One property can be nominated as the badge: its value is then drawn as a small chip on the shape itself.

Shape data is stored under the node's meta.data, travels with copy/paste and JSON export, and is visible to the AI layer and to the MCP server (get_node_data / set_node_data).

Tables#

Data ▸ Table edits the selected table node: header row, cell values, add and remove rows and columns (also on the right-click menu and in the palette under Table). The canvas overlay lets you drag column widths directly.

Conditional formatting#

Data ▸ Conditional formatting holds a list of rules, evaluated in order:

if <property> <operator> <value>  then <style>

Operators: equals, is not, >, ≥, <, ≤, contains, starts with, ends with, is empty, is not empty, is true, is false, is before, is after — the comparison ones are offered only for the types they suit.

Rules are re-evaluated when you edit shape data, edit a rule, or run Apply conditional formatting from the palette.

Manual overrides are never clobbered irreversibly. Whatever a rule wrote last time is recorded on the node; re-applying first lifts exactly those values, and only the ones you have not since changed by hand. Change a fill manually afterwards and the rule leaves it alone; delete the rule and the shape goes back to what it was.

CSV and JSON in#

Data ▸ Import & export turns a table of records into shapes.

  1. Paste or choose a CSV/JSON file. The delimiter is sniffed (comma, semicolon, tab or pipe); quoted fields, embedded newlines, CRLF, a BOM and ragged rows are all handled.
  2. Map columns to id, text, body, kind, style and parent — or, in edge mode, to from, to and label.
  3. Pick a layout: grid, tree or rows.
  4. The panel previews exactly what will be inserted; inserting is one undo step.

Unmapped columns become shape data on the node they came from, with their type guessed.

CSV out#

Palette → Export node data as CSV writes one row per node — id, kind, text, body, x, y, w, h, parent followed by one column per shape-data key found anywhere in the document — so a diagram can go out to a spreadsheet and come back.