TableGripSchema & operations
Schema tools
Design tables, compare schemas and generate migrations, draw ER diagrams, import data and generate test rows.
The schema tools live in the explorer's right-click menus:
| Right-click a… | You get |
|---|---|
| Schema (or a connection, for its default schema) | New Table…, Import Data…, Show ER Diagram, Compare Schema With… |
| Table | Edit Structure…, Show in ER Diagram, Import Data…, Generate Test Data… |
#How changes are applied
Every tool that changes the schema or writes rows works the same way:
- It shows the statements first, so you see exactly what will run.
- It respects read-only connections (the tools that write are hidden there) and asks again on Production.
- On databases with transactional DDL (PostgreSQL, SQLite, SQL Server, DuckDB) everything runs all-or-nothing. Elsewhere, statements apply one by one and stop at the first failure; the dialog tells you this up front.
#Table designer
Open it with New Table…, Edit Structure…, or the pencil in a table's Structure view.
- Columns, indexes and foreign keys are editable lists.
- Renames are renames, not a drop and re-add, so no data is lost, and they follow into indexes and keys.
- The SQL preview updates as you type. Press ⌘ S to apply.
On SQLite, some changes (such as changing a column's type) can't be made in place. The designer then rebuilds the
table the way SQLite documents it: rows, indexes, triggers, UNIQUE and CHECK constraints come back, and foreign keys
are checked before committing.
#Schema compare
Compare Schema With… compares any two relational connections and schemas, for example staging against production.
- Differences are listed per table, each with a checkbox. Drops start unchecked, so nothing is removed unless you ask for it.
- The migration is written in the target's dialect.
- Copy it, save it, open it in a console, or run it through the usual confirmation dialog.
#ER diagram
Show ER Diagram draws a schema; Show in ER Diagram draws a table and its neighbours.
- Tables are laid out in layers by their foreign keys, with crow's-foot ends.
- Drag tables to arrange them. Pinch or ⌘-scroll to zoom; scroll with two fingers to pan.
- Find a table by name, and double-click one to open its rows.
- Export as PNG (light or dark, at 2×) or as a Mermaid
erDiagramto paste into docs or a pull request.
#Import data
Import Data… loads a file into a new or existing table.
| Format | Notes |
|---|---|
| CSV / TSV | Delimiter and encoding detected |
| Excel | .xlsx workbooks |
| JSON | An array of objects, or NDJSON |
- Preview the rows first.
- Choose a new table (column types are inferred, names cleaned to
snake_case) or an existing table (columns matched by name). - Rows are inserted in batches in one transaction. If a row fails, it's named by its line number and nothing is kept. A table created for the import is dropped again.
Tip
To work with a file without importing it at all, open it directly as a table. See CSV, Excel & JSON files.
#Generate test data
Generate Test Data… fills a table with realistic rows:
- A generator per column, picked from its name and type: emails that match the name, phone numbers, prices, past timestamps and more. Change any of them.
- NULL ratios and uniqueness per column.
- Foreign keys are drawn from the referenced table's existing rows, so the data stays consistent.
- The same seed gives the same rows, which makes test setups reproducible.
- Preview before inserting.