Turns a reporting event (tfl_read_ars_json(), or a tfl_ars()) into
the specs it says:
Arguments
- ars
A
tfl_ars()ortfl_read_ars_json().- table
Also a table spec.
Value
A list: ard (a tfl_ard_spec()), report (a tfl_table_spec()
holding the report sheets) and, with table = TRUE, table.
Attribute unmapped: a data frame where, item, reason.
Details
the ARD spec: analysis sets as
populations, analyses asanalysesrows (method, dataset, population,where,by,variables,statistics,purpose,reason), the datasets they read asdatasets(without a path: ARS does not say where the ADaM is);the report spec: one
reportrow per output in the list of contents' order, its file, and its display's titles, footnotes, header and footer;with
table = TRUE, a table spec with thevariablesrows the groupings give: a grouping's listed groups as the variable'slevels.
tfl_ars()'s own layout comes back as it was written (one row with
several variables, a hierarchy as one row); anyone else's analysis
becomes one row each. A method is read from tflspec's ids, else from
what its operations compute; an analysis whose method tflspec has no
keyword for (an ANOVA, SAS code only ...) is not made a row and is
listed in unmapped.