Schema Explorer

The navigator is a tree of each connected database's objects, shaped like the engine's own hierarchy and loaded as you expand it, so a large catalog never blocks the first expand. Below the tree, an inspector shows the selected object's detail.

Engine-Shaped Hierarchy

• PostgreSQL, TimescaleDB and CockroachDB: Databases > database > Schemas > schema > folders for Aggregates, Collations, Foreign Tables, Functions, Materialized Views, Operators, Procedures, Sequences, Tables, Trigger Functions, Types and Views. Beside Databases sit Login/Group Roles, on PostgreSQL and TimescaleDB Tablespaces, and a Server node holding Server Info, Sessions and Locks. A table expands to Columns, Indexes, Foreign Keys and, except on CockroachDB, Triggers.
• MySQL / MariaDB: Databases (here a schema is a database) > Tables, Views, Stored Procedures and Functions; a table expands to Columns, Indexes, Foreign Keys and Triggers. A Server node holds Server Info, Sessions, Users and Roles.
• SQL Server: Databases > database > Tables, Views, Synonyms, Programmability (Stored Procedures, Functions, Database Triggers, Assemblies, Types, Sequences), Storage (Filegroups, Log Files) and Security (Users, Roles, Schemas), with each object named schema.name; beside Databases, Security > Logins and a Server node holding Server Info and Sessions. A table expands to Columns, Indexes, Foreign Keys and Triggers.
• Oracle: Schemas (Oracle schemas are users) > Tables, Views, Procedures, Functions, Sequences and Types. A Server node holds Server Info, Users, Roles and Tablespaces.
• SQLite: no server or namespace level (single file): Tables, Views and Indexes sit directly under the connection; a table expands to Columns, Indexes, Foreign Keys and Triggers.
• MongoDB: Databases > Collections and Views; a collection expands to Fields, sampled from its first 100 documents, and Indexes. A Server node holds Server Info, Users and Roles.
• Apache Cassandra: Keyspaces > Tables, Materialized Views and Functions, with columns, and a Server node whose Server Info shows the release, cluster name and data centre.
• Redis, Neo4j and Elasticsearch: key namespaces with their key counts, node labels and relationship types, and indices with their document count, size and health; each has a Server node with Server Info.

Under a Databases level, your own databases sit at the top and the system databases are gathered into a System Databases folder (PostgreSQL postgres/template0/template1; SQL Server master/model/msdb/tempdb; MySQL information_schema/mysql/performance_schema/sys; MongoDB admin/config/local) so operational objects stay out of the way until needed.

Server-Level Introspection

Where an engine has a Server node, Server Info shows the server's version and, where the engine reports them, its uptime and main settings; Sessions lists the server's other sessions with their user, application, state and current statement (your own browsing session is left out); Locks lists the locks held and awaited on PostgreSQL and TimescaleDB, and on CockroachDB the row locks some transaction is waiting for, and for a waiting lock names the session or transaction holding it; and Users, Roles and Tablespaces list the server's accounts and storage.

Object Detail & Definitions

Selecting a table opens the inspector with columns (type, nullability, default, constraints, comments), indexes (name, columns, method, uniqueness), constraints (PK, FK with referenced table/column, unique, check), and, where reported, statistics-based approximate row count and on-disk size, no full COUNT(*) scan. Column rows carry type, nullability, default expression (e.g. sequence nextval), and PK/FK/UQ role markers. The inspector offers only the tabs the engine can answer: all eight (Columns, Indexes, Foreign Keys, Constraints, Triggers, Partitions, Statistics, Grants) on PostgreSQL, TimescaleDB and Oracle; all but Triggers on CockroachDB; all but Partitions on MySQL, MariaDB and SQL Server; Columns, Indexes, Foreign Keys, Triggers and Statistics on SQLite; Columns, Indexes and Statistics on MongoDB; Columns on Cassandra, ClickHouse and Redshift. Hovering a row in the navigator names what the object is, then whatever the engine reports for it, such as its approximate row count, size, owner, return type or comment; folders show nothing on hover. A PostgreSQL table that has not been analyzed yet, and every view, shows no row count.

Search

The Search objects box searches every connected connection once you type two characters and lists up to 50 matches per connection. It matches part of a name: tables, views, materialized views and sequences (and their comments) in the connected database on PostgreSQL, TimescaleDB and CockroachDB; tables and views (and their comments) in every database on MySQL and MariaDB; tables, views, procedures and functions in the connected database on SQL Server; tables, views, procedures and functions you can see on Oracle, which compares the name in upper case; tables and views on SQLite; and collections in every database on MongoDB. The engines' own catalogs are left out, including CockroachDB's crdb_internal and TimescaleDB's internal schemas. Apache Cassandra, Redis, Neo4j and Elasticsearch objects are not searched: hovering the box names those connections, and when they are the only ones connected the results say so. If a connection's search fails, its name and the error show under the box while the other connections' matches still list. Clearing the box restores the tree.

Context Actions

Right-clicking a table/view/collection offers:

• Select top 100 rows and Select top 1000 rows (Show first 100 documents and Show first 1000 documents on a MongoDB collection) retarget the editor to that object's connection and database (the toolbar's database switches to it), write a query for its engine, and run it there: LIMIT (PostgreSQL/MySQL/SQLite), SELECT TOP (SQL Server), FETCH FIRST n ROWS ONLY (Oracle), or a JSON find with a limit (MongoDB), in the form the IDE runs MongoDB queries.
• Script definition to editor puts the object's CREATE statement in the editor, read from its own database. It is offered for tables, views, materialized views, functions, aggregates, procedures, sequences, triggers and indexes on PostgreSQL, TimescaleDB and CockroachDB; tables, views, functions and procedures on MySQL, MariaDB and SQL Server; tables, views, functions, procedures and sequences on Oracle; tables, views, indexes and triggers on SQLite; tables and views on ClickHouse; and not on engines that have no DDL to show.
• Script first 1000 rows as INSERT (Script first 1000 documents as insertOne on a MongoDB collection) reads up to 1000 rows of a table and puts the insert script for them in the editor, in the engine's own language as described under Export Results. It is offered for tables on the SQL engines and Cassandra and for MongoDB collections; a table with no rows says so instead.
• Insert into editor adds the object's name, qualified by its schema, on a new line at the end of the editor's text without running anything.
• Add to favorites pins the object into a Favorites section, kept in this browser, for fast access across a large catalog.
• Copy Name copies the qualified name (e.g. public.users) to the clipboard.

Introspection Sources

Depth comes from querying each engine's own catalog rather than a normalized abstraction: information_schema + pg_catalog (PostgreSQL/Cockroach/Timescale), information_schema (MySQL/MariaDB), INFORMATION_SCHEMA + sys views (SQL Server), the ALL_* dictionary views (Oracle), sqlite_master + PRAGMA table_info() (SQLite), listCollections + document sampling (MongoDB), and the key names, read in batches with SCAN so Redis keeps serving its other clients, grouped by prefix (Redis). On a Redis holding more than 100,000 keys the counts cover the first 100,000 read, and hovering a namespace says so.