Data grid & editing

Every result — a query, a table preview, a Mongo collection — lands in the same grid: a custom virtualized grid that renders only the visible window, so scrolling stays smooth whether a table has a thousand rows or millions.

Browsing without SQL#

Click a table in the Explorer to open it directly — no SELECT needed. The filter drawer narrows rows by column conditions without hand-writing a WHERE, and the table's structure view shows column types, keys, and indexes. Large tables paginate their previews; for very large ones, query with an explicit LIMIT in the SQL editor.

Inline editing#

• Double-click a cell to edit it in place. NULL handling, JSON columns (rendered and edited as structured JSON rather than a string blob), dates, and binary values all edit as their real types.
• Stage, then commit — edit multiple cells across multiple rows; staged changes are highlighted until you commit them all in one atomic transaction with ⌘ S / Ctrl+S. Nothing is written until you say so.
• Row inspector — open a single row in a side panel to see every column's full contents without horizontal scrolling.
• Insert and delete rows straight from the grid.

Row & document compare#

Select two or more rows — grid rows in any SQL connection, or documents in MongoDB — and compare them field by field: the union of every field seen across the selection, missing fields included. Toggle only show differences to hide identical fields, and copy the whole comparison as JSON in one click for a PR, ticket, or chat.

Export#

Grid results export to the usual formats (CSV, JSON, SQL) from the grid toolbar — handy for handing data to someone without database access.

Troubleshooting#

• SQLite database is locked — SQLite allows one writer at a time; wait for other writers or close external editors before committing.
• Grid shows NULL for a type — exotic types decode as the engine prints them (Postgres interval, inet, money, arrays, JSONB and friends); if one still reads wrong, report it.