Skip to contents

Connect to Databricks SQL Warehouse

Source: R/databricks-dbi.R
dbConnect-DatabricksDriver-method.Rd

Connect to Databricks SQL Warehouse

Usage

# S4 method for class 'DatabricksDriver'
dbConnect(
  drv,
  warehouse_id = NULL,
  http_path = NULL,
  catalog = NULL,
  schema = NULL,
  staging_volume = NULL,
  disposition = c("EXTERNAL_LINKS", "INLINE"),
  max_active_connections = 30,
  fetch_timeout = 300,
  show_progress = TRUE,
  token = db_token(),
  host = db_host(),
  ...,
  query_tags = NULL
)

Arguments

drv

A DatabricksDriver object

warehouse_id

Optional ID of the SQL warehouse to connect to

http_path

Optional HTTP path for the SQL warehouse; if provided, the warehouse ID is extracted from this path

catalog

Optional catalog name to use as default

schema

Optional schema name to use as default

staging_volume

Optional volume path for large dataset staging

disposition

Query disposition mode to use by default for DBI query results. Use "EXTERNAL_LINKS" for large results or "INLINE" for smaller results that must avoid direct cloud-storage result downloads.

max_active_connections

Maximum number of concurrent download connections when fetching query results (default: 30)

fetch_timeout

Timeout in seconds for downloading each result chunk (default: 300)

show_progress

If TRUE, show progress updates by default for DBI queries, dbplyr collection, and table writes (default: TRUE)

token

Authentication token (defaults to db_token())

host

Databricks workspace host (defaults to db_host())

...

Additional arguments (ignored)

query_tags

Optional named list of execution-level query tags, for example list(team = "informatics", environment = "dev"). Supply at most 20 tags with unique, non-empty names and single string values. Use a NULL value for a key-only tag. NULL or list() omits tags from the request. Tags are recorded in system.query.history.query_tags; they are distinct from the saved-query metadata tags in db_query_create().

Value

A DatabricksConnection object

Details

Provide either warehouse_id or http_path. When http_path is supplied, the warehouse ID is extracted from the /warehouses/<id> segment. Connections are stateless: connection-level query_tags are sent with each statement, including connection validation, metadata queries, and table writes. They do not create a server-side SQL session.

See also

Other DBI Methods: dbExecute,DatabricksConnection,character-method, dbGetQuery,DatabricksConnection,character-method, dbSendQuery,DatabricksConnection,character-method, dbSendStatement,DatabricksConnection,character-method