Set up DuckDB connection with views for available data types — accessParquetData • parkinsonsMetagenomicData

'accessParquetData' is a wrapper function for 'db_connect' and 'retrieve_views'. A DuckDB connection is established and views are created for either all provided local files or all data types available in a repo of interest (see inst/extdata/parquet_repos.csv). When using a remote repo, a vector of specific data types can be supplied as doing this for all data types can take longer.

Usage

accessParquetData(
  dbdir = ":memory:",
  repo = NULL,
  local_files = NULL,
  data_types = NULL
)

Arguments

dbdir: Location for database files. Should be a path to an existing directory in the file system or the value ':memory:' to keep data in RAM. Default: ':memory:'
repo: String (optional): Hugging Face repo where the parquet files are stored. If NULL and local_files is also NULL, the repo listed as the default in get_repo_info() will be selected. Default: NULL
local_files: String or vector of strings (optional): path(s) to parquet file(s). Default: NULL
data_types: Character vector (optional): when using a remote repo, a list of data types to establish database views for. If NULL, views will be created for all available data types. Default: NULL

Value

DuckDB connection object of class 'duckdb_connection'

Details

Files stored remotely and locally cannot be combined in the same connection.

Examples

if (FALSE) { # \dontrun{
if(interactive()){
 prepared_db <- accessParquetData()
 DBI::dbListTables(prepared_db)

 single_type <- accessParquetData(data_types = "pathcoverage_unstratified")
 DBI::dbListTables(single_type)
 }
} # }