Unpack response JSON columns — unpack_response

This function unpacks response JSON payloads that are distributed across several columns, for example question_0_content, sums_content, or responses. Matching timestamp columns are detected by replacing a _content suffix with _ts, for example question_0_content is matched to question_0_ts. This is particularly useful for BKT-like question-slot preparation where coded responses are distributed across question_*_content columns instead of one coded column.

Usage

unpack_response_jsons(
  responses,
  response_cols = NULL,
  id_cols = NULL,
  keep_empty_payloads = FALSE,
  keep_empty_rows = TRUE,
  progress = TRUE
)

Arguments

responses: Data frame. Response data with one or more JSON columns.
response_cols: Character vector. Columns to unpack. If NULL, columns are auto-detected by parsing non-empty JSON payloads that contain response record fields such as id, status, value, subform, code, and score.
id_cols: Character vector. Columns to keep as identifiers. If NULL, all non-JSON and non-timestamp columns are kept.
keep_empty_payloads: Logical. Whether missing, empty, and [] payloads should be kept as rows with missing response identifiers.
keep_empty_rows: Logical. Whether source rows that do not produce any unpacked JSON records should be kept as one row with missing response fields.
progress: Logical. Whether to show a progress bar while parsing JSON payloads.

Value

A tibble with the identifier columns, response slot metadata, and the parsed JSON fields.