ClickHouse Action Block

What it does: Runs SQL against a ClickHouse database over its HTTP interface — analytical queries, bulk inserts, and schema management — and returns structured rows to your workflow.

⚡

In simple terms: Point the block at your ClickHouse Cloud or self-hosted endpoint, write SQL, and get back rows, columns, and a rowCount you can feed into an LLM, a report, or another block.

Credentials

Field	Required	Description
HTTP URL	Yes	HTTP interface endpoint, e.g. `https://abc.us-east-1.aws.clickhouse.cloud:8443`. The protocol defaults to `https` if you omit it. Self-hosted instances usually listen on port `8123`.
Username	No	ClickHouse user (often `default`). Sent as the `X-ClickHouse-User` header.
Password	No	Sent as the `X-ClickHouse-Key` header.
Default database	No	Used whenever a block does not set its own Database field.

In ClickHouse Cloud, find these under your service → Connect → HTTPS.

Actions

Action	What it does	Output
Run Query	Executes any SQL statement with optional server-side parameters	`{ rows, columns, rowCount, statistics, raw }`
Insert Rows	Inserts a JSON array of objects via `INSERT INTO … FORMAT JSONEachRow`	`{ inserted, table, raw }`
List Databases	`SHOW DATABASES`	`{ rows, columns, rowCount }`
List Tables	`SHOW TABLES FROM <database>`	`{ rows, columns, rowCount }`
Describe Table	`DESCRIBE TABLE <database>.<table>`	`{ rows, columns, rowCount }`
Create Table	Runs your raw DDL, or builds a `CREATE TABLE` from table + schema fields + engine + order by	`{ created, statement, raw }`
Truncate Table	`TRUNCATE TABLE IF EXISTS` — requires confirmation	`{ table, statement, raw }`
Drop Table	`DROP TABLE IF EXISTS` — requires confirmation	`{ table, statement, raw }`
Ping	`GET /ping` to verify connectivity	`{ ok, status, response }`

Fields

Field	Used by	Description
`query`	Run Query	The SQL statement.
`params`	Run Query	JSON object of parameters, sent as `param_<name>`.
`database`	most actions	Database name — falls back to the credential's default.
`table`	Insert / Describe / Create / Truncate / Drop	Table name, optionally `database.table`.
`rows`	Insert Rows	JSON array of row objects.
`ddl`	Create Table	Raw `CREATE TABLE` statement (takes precedence).
`schemaFields`	Create Table	`{"id":"UInt64","name":"String"}` or `[{ "name": "id", "type": "UInt64" }]`.
`engine`	Create Table	Table engine, defaults to `MergeTree`.
`orderBy`	Create Table	`ORDER BY` expression, defaults to `tuple()`.
`confirm`	Truncate / Drop	Must exactly match the table name.
`maxRows`	Run Query	Caps the result set (`max_result_rows`).
`format`	Run Query	ClickHouse output format, defaults to `JSON`.

Query parameters

ClickHouse binds parameters server-side, which keeps interpolated workflow variables out of the SQL text:

SELECT event, count() AS total
FROM events
WHERE user_id = {uid:UInt64} AND event = {name:String}
GROUP BY event

With Params:

{ "uid": "42", "name": "purchase" }

Tips

Database and table names are validated: only letters, digits and underscores (plus one dot for database.table) are accepted. query and ddl are intentionally raw — they are your own SQL.
Truncate and Drop refuse to run unless Confirm matches the table name exactly.
Setting Format to something other than JSON (e.g. CSV, TSV) returns the response body in raw instead of parsed rows.
Use Max Rows on exploratory queries so a wide SELECT * cannot flood the workflow output.

Chargebee CoinGecko