For everyone

Running queries#

The SQL editor is where you write and run Trino SQL against a cluster.

Choose your context#

At the top of the editor, pick three things:

  • Cluster — the Trino cluster that will execute your query. It must be Running (see Understanding clusters).
  • Catalog — the data source, for example tpch, tpcds, or any catalog your admin configured (S3 + Glue, a database like PostgreSQL or Snowflake, MongoDB, Elasticsearch, and more).
  • Schema — the schema within that catalog, for example sf1.

The schema browser on the left lists loaded catalogs, schemas, and tables. Click a table to insert its fully-qualified name into the editor. The browser and autocomplete are backed by a server-side metadata cache, so they keep working from the last known schema even when the cluster is suspended.

Write and run#

Type your SQL in the editor. It has syntax highlighting, line numbers, and autocomplete that's aware of your selected context (catalogs, schemas, tables, columns, keywords, and functions).

There are three run modes:

  • Current — run the statement your cursor is in.
  • Selected — run only the text you've highlighted.
  • All — run every statement, separated by semicolons, in order.

Click Run (or press Ctrl/Cmd + Enter). While a query runs you'll see its status, elapsed time, and row count update live. Click Cancel to stop a running query.

Click Explain to see the query plan without executing it — TrinoHub wraps your statement in EXPLAIN and shows the result.

Results, charts, and profile#

The results card has three tabs:

  • Results — your rows in a table.
  • Chart — pick a chart type (bar, line, area, or pie) and the X and Y columns to visualize the result set.
  • Query profile — the Trino query ID, timing, context, and row counts, plus, for a query that reached a running coordinator, an execution detail breakdown: coordinator state, queued/execution/CPU time, peak memory, splits, and a per-stage table.

Result limits: the browser shows up to 1,000 rows or 10 MB, whichever comes first. When a result is capped you'll see a notice. To get more rows, export to CSV.

Cached results#

Re-running an identical read-only query (same cluster, catalog, schema, and SQL — ignoring whitespace, comments, and keyword case) within the result-cache window serves the stored result instantly instead of contacting the cluster. A cache hit even skips resuming an auto-suspended cluster.

When a result comes from the cache the results card shows a Cached — ran N minutes ago badge, and the run is marked Cached in query history. Cached results are point-in-time snapshots; click Run fresh to bypass the cache and re-execute.

Cache entries are per user — one user's cached results are never served to another. Admins set the cache window (default 10 minutes) in Settings → Result cache; 0 disables it. Results expire after the window and nothing larger than the normal result caps is ever stored.

The cache steps aside whenever freshness or access could be in question:

  • Running a write statement invalidates your own cached entries on that cluster, so a follow-up SELECT re-executes.
  • Admin changes to roles, policies, tags, users, catalogs, or clusters flush the whole cache.
  • A disabled, failed, or manually suspended cluster never serves cached results — only Running and auto-suspended clusters do.
  • Scheduled jobs always execute for real; they never consume the cache.

Exporting to CSV#

Click CSV to download the result. CSV export streams a larger capped buffer of 10,000 rows or 50 MB. Use Run CSV to run and download in one step. CSV export of a cached run reuses the stored buffer — it never re-executes the query.

Saving, sharing, and reusing queries#

  • Save stores the current query in your Saved queries library (search, sort, and reopen from the right rail).
  • Share a saved query with a role at one of three levels — view, view and run, or edit. Shared queries appear in recipients' libraries labeled with the owner and their access level; only the owner can delete a shared query.
  • The Recent queries rail and the Query history view list what you've run — open any entry to load it back into the editor.

Tips#

  • Find / Replace — search and edit within the editor.
  • Format — tidy up common SELECT clauses.
  • Commands — open the command palette for editor tools and actions.
  • Global search — press Ctrl/Cmd + K anywhere to search across clusters, catalogs, cached tables, saved queries, and notebooks; results are filtered to what your roles allow.
Edit this page on GitHub