Open a JSON dataset as a lazy dplyr tbl

Validates the URL, loads the azure extension, then returns a lazy dplyr::tbl() over the dataset. Use az_conn() first if the connection needs Azure extensions, settings, or secrets.

Usage

tbl_json(conn, url, name = NULL, replace = TRUE, ...)

Arguments

conn: A DuckDB connection.
url: Character scalar. Azure Blob URL. Supports glob patterns.
name: Optional character scalar. Name to register the view under in DuckDB. When NULL (default) the dataset is scanned directly.
replace: Logical. Replace an existing view of the same name. Default TRUE. Ignored when name = NULL.
...: Reader options forwarded to DuckDB's read_json_auto().

Value

A dplyr::tbl() backed by the JSON dataset.

Details

When name is NULL the dataset is queried directly via read_json_auto() with no persistent object registered on the connection. When name is supplied the dataset is first registered as a VIEW via load_json(), then referenced by name.

Examples

if (FALSE) { # \dontrun{
# Requires a live Azure account, credentials, and network access.
conn <- az_conn()
tbl_json(conn, "abfss://container@account/data/*.json") |>
  dplyr::collect()
} # }