Result metadata, one row per component result

Calls POST /results/metadata with the ids as a JSON body, so any number of ids fits in one request. Results the token cannot see are left out by the server without an error. The answer is flattened with jatos_flatten_metadata(); see there for the columns.

Usage

jatos_results_metadata(
  study_id = NULL,
  batch_id = NULL,
  component_id = NULL,
  study_result_id = NULL,
  component_result_id = NULL,
  group_id = NULL,
  conn = jatos_connection()
)

Arguments

study_id, component_id

Ids (integer) or uuids (string) to select results by, one kind per argument; NULL (the default) means no restriction on that dimension. Uuids go to the server as studyUuids and componentUuids.

batch_id, study_result_id, component_result_id, group_id

Ids to select results by. NULL (the default) means no restriction on that dimension. The server takes no uuid for these.

conn

A jatos_connection().

Value

A metadata tibble, one row per component result, with the columns documented in jatos_flatten_metadata().

Details

Every id argument accepts one or more integer ids; study_id and component_id also accept uuid strings, which survive a study's export and import on another server while its id changes. At least one must be given; several are combined as the server combines them (the union of everything selected). A message reports the host, the number of study results and component results, and how many component results are not FINISHED; silence it with options(rlib_message_verbosity = "quiet") or suppressMessages().

See also

jatos_study_results() for one row per study result, jatos_read_metadata() to rebuild the same tibble from a local cache.

Examples

if (FALSE) { # \dontrun{
# needs a JATOS server and a stored API token
meta <- jatos_results_metadata(study_id = 12)
meta[meta$component_state == "FINISHED" & meta$data_size > 0, ]
} # }