Databases, tables and structure

The sidebar of a connection tab is where you pick a database and a table; the Structure tab is where you change them. The words follow the engine — collection on MongoDB, index on Elasticsearch, file for a SQLite folder — and Redis has keys instead of tables (see Redis and Valkey).

The sidebar header

The connection’s engine logo and name, its PROD or READ-ONLY badge, the server version, and Disconnect (closes the connection and any SSH tunnel).

The database picker

Click the database name, or press Database → Choose Database… (mod+P).

• Search — Search N databases; the count shows matches (3 of 124) and the matched part is bold.
• Keys — ↑ / ↓, PgUp / PgDn, enter to open, Esc to close.
• ★ Star — Star: keep at the top. Starred and Recent databases are grouped first, per connection.
• Your databases come next, then System — the server’s own, with a built in badge, folded when there are many.

Next to the picker:

• ↻ — Reload the databases and tables (mod+R, Database → Refresh).
• + — create a database (below). Not for Redis, Valkey or Elasticsearch, and off on a Read-only connection.

On Redis and Valkey the picker lists db0 … dbN with their key counts. On Elasticsearch the “database” is the cluster itself. A SQLite folder lists each file as a database.

The table list

• Filter tables narrows the list.
• The count’s tooltip says N of your own, M built in.
• + New creates a table.
• Each row shows the table’s rows, short (1.2k); hover for rows and size.
• Views and materialized views are listed as views. PostgreSQL tables outside public show as schema.table.

Your own tables come first. System tables are grouped apart, folded, with a built in badge: information_schema, performance_schema and sys, the server’s tables in mysql, sqlite_* tables, MongoDB’s system.* collections and Elasticsearch’s dot indices.

The table menu

Right-click a table:

• Show Data (mod+1) · Show Structure (mod+2)
• Copy Name · Copy CREATE Statement
• Export Table…
• Rename Table… — the Rename <table> prompt asks for a New name. MongoDB renames the collection.
• Duplicate Table… (Duplicate Collection… / Duplicate Index… on MongoDB and Elasticsearch; not for views) — Duplicate <table> asks for a New name (<table>_copy to start with) and has Copy the rows too, on by default. SQL tables keep their columns, keys and indexes; foreign keys and triggers stay with the original. An Elasticsearch index keeps its mappings and analysis settings, a MongoDB collection its indexes. On a Production connection it asks first, like any write.
• Truncate Table… — not for views.
• Drop Table…

The same actions are in the Table menu of the menu bar. Truncate and drop ask you to type the table’s name — see Dropping and emptying.

Select several tables

• ⌘-clickCtrl-click adds or removes a table.
• ⇧-clickShift-click picks a range.
• mod+A (Edit → Select All, with the list focused) picks every table shown.
• Esc clears the selection.

A bar appears: N selected · right-click for actions, with ✕ to clear.

Bulk actions on several tables

Right-click the selection for the menu titled N Tables Selected:

On a Read-only connection the writes are greyed out, marked Read-only connection.

Truncate and drop go child-first, so tables that point at others are handled before the tables they point at. You type the database’s name to confirm — or the connection’s name on Production.

Force — disable integrity checks lets them go through when foreign keys are in the way:

Without Force, a table outside your pick that points at one of them makes DBCatz refuse.

The sidebar foot

Above the view sits the breadcrumb — db › table and the row count — and the view tabs Data · Structure · Query · Logs (Redis / Valkey: Keys · Command · Logs).

The Structure tab

Table → Show Structure (mod+2). The sections depend on the engine.

Columns

SQL engines show Columns. MongoDB shows Fields (from a sample of documents) — read from the first 100 documents, since a collection has no fixed columns. Elasticsearch shows the index’s Mapping.

Add or change a column

+ Column adds one; Change edits one — including its name, since renaming happens there.

• Name.
• Type — a type picker: a grouped dropdown of the engine’s types, then length or precision and scale, the values of an ENUM or SET, and modifiers such as UNSIGNED.
• Default — empty = none; expressions like CURRENT_TIMESTAMP work.
• Allows NULL.
• Auto increment — when adding.
• Character set, Collation and Comment.
• Unique.

Every change goes through the structure checks before the ALTER runs.

Per engine:

• SQLite — changing an existing column is not possible: the dialog says SQLite cannot change a column. Its collations are BINARY, NOCASE and RTRIM, and it has no column comments.
• PostgreSQL comments are set with COMMENT ON COLUMN.

Column display order

Drag ⠿ to set the order columns are shown in, here and in the Data grid. It is a view setting for you — not an ALTER; the table is untouched. Reset order puts it back.

Indexes

For SQL engines and MongoDB: Name, Columns and Kind (primary, unique, index), with Drop primary key or Drop index.

+ Index opens Add index: a Name (optional), Columns (in order) with + Column, and Unique.

Foreign keys

For SQL engines: Name, Columns, References, On delete and On update. This list is read-only — foreign keys cannot be added from the Structure tab; write the ALTER TABLE … ADD FOREIGN KEY in the Query tab.

CREATE statement

The table’s CREATE statement (SQL) or its Definition (other engines), with Copy.

Create a table

Table → Create Table… (mod+shift+T), + New in the table list, or Create table on an empty database. The dialog is Create

Each column row has Name, Type, Null, Default, PK, Auto (auto increment), UQ, ⋯ for character set, collation and comment, and ✕ to remove it. Below: + Column, and + Timestamps (SQL engines) to add created_at and updated_at columns in one go.

• Elasticsearch — the fields become the new index’s mapping; leave them empty for a dynamic mapping.
• MongoDB — just a name: A collection has no fixed columns.

Create a database

Database → Create Database… (mod+shift+N) or + next to the picker. Give it a Name, and:

• MySQL / MariaDB — Character set and Collation, utf8mb4 / utf8mb4_unicode_ci by default. The hints tell you which collations are case-insensitive, binary or case-sensitive.
• PostgreSQL — Encoding and Collation (server default; it sets LC_COLLATE and LC_CTYPE). Values other than the default use template0.
• MongoDB — a database exists once it has a collection.

There is no Create database on Redis, Valkey or Elasticsearch. Database → Drop Database… asks you to type the name.

Last updated 2026-11-03