Stated Fallbacks

Review ask

As a clinical trial statistician or programmer, challenge anything in this benchmark: describe a problem in Comments, or propose a fix with the Edit button on any source file.

Task: Would a real study need this derivation, specified this way? Does it follow CDISC ADaM: variable names, controlled terms, record structure? Flag any rule that is ambiguous or that a sponsor would handle differently.
Inputs: Do they look like data collected in a real trial, and do they include the situations you would expect to meet in one?
Expected output: Would independent QC programming reproduce every value from these inputs and rules?

The other tabs beside Comments are background for assessment, not under review.

Summary

EditUnder reviewWhat this benchmark means

README.md

Goal: derive vital-signs analysis values where each fragile read states its fallback beside it: an unmapped or absent test code, a collected result that cannot be read as a whole number, and a subject with more than one dosing record.

Input: one spec.yaml. VS carries the vital-signs records: subject, sequence, test code, the result as entered text, and units. EX carries the dosing records: subject, sequence, treatment, dose, and start date.

Fallbacks:

PARAM codes each test: a collected code with no dictionary entry, and a record with no collected code at all, both take the stated "Other measure" label.
AVAL reads the collected result as a whole number: a result that cannot be read that way, such as "ND" or the fractional "98.6", is left blank by the stated fallback. A record with no collected result is blank because there is nothing to read.
LASTDOSE takes each subject's latest dose: a dosing record with no dose never qualifies, the dosed record with the latest start date wins, and a subject with no dosed record keeps the dose blank.

Note: each fallback is stated beside the read it guards, so the run completes with the stated answers instead of failing; a read whose condition never occurs keeps its ordinary value.

YAML specification

1 spec file

spec.yaml

Edit72 lines

1schema_version: "1.0"2domain: ADVS3keys: [USUBJID, VSSEQ]4input:5  VS: {path: input/vs.csv, types: {VSSTRESN: str}}6  EX: {path: input/ex.csv, types: {EXDOSE: float}}7base: VS89output:10  path: advs.csv11  columns: [STUDYID, USUBJID, VSSEQ, PARAM, AVAL, LASTDOSE]1213columns:14  - name: STUDYID15    type: str16    label: Study Identifier17    derivation: VS.STUDYID1819  - name: USUBJID20    type: str21    label: Unique Subject Identifier22    derivation: VS.USUBJID2324  - name: VSSEQ25    type: int26    label: Sequence Number27    derivation: VS.VSSEQ2829  # A collected code with no dictionary entry, and a record with no30  # collected code at all, both take the stated fallback label.31  - name: PARAM32    type: str33    label: Parameter34    derivation:35      mapping:36        source: VS.VSTESTCD37        dict:38          SYSBP: Systolic Blood Pressure39          DIABP: Diastolic Blood Pressure40          PULSE: Pulse Rate41        missing: Other measure42        unmapped: Other measure4344  # The collected result is entered as text. "ND" and the fractional45  # "98.6" cannot be read as whole numbers, so the stated fallback46  # leaves them blank; a record with no collected result is blank47  # because there is nothing to read. This `unconvertible` sits beside48  # `value` and answers only a failed conversion to int.49  - name: AVAL50    type: int51    label: Analysis Value52    derivation:53      value:54        source: VS.VSSTRESN55      unconvertible: null5657  # A dosing record with no dose never qualifies; among the dosed58  # records the later start date wins. A subject with no dosing59  # record keeps the dose blank. Multiple matches fail by default;60  # `order_by` + `keep` selects the later start date instead. The61  # stated `absent` answers a variable absent from context; every62  # row here reads a present variable, so no row takes it.63  - name: LASTDOSE64    type: float65    label: Latest Dose66    derivation:67      source:68        variable: EX.EXDOSE69        filter: "EX.EXDOSE IS NOT NULL"70        order_by: [EX.EXSTDTC]71        keep: last72        absent: null

Inputs

Under review2 source files

input/ex.csv

Edit4 rows

input/ex.csv
STUDYID	EXSEQ	EXTRT	EXDOSE	EXSTDTC
YAMAA-01	1	DRUG A	10	2026-01-05
YAMAA-01	2	DRUG A	20	2026-02-05
YAMAA-01	3	DRUG A	—	2026-03-05
YAMAA-01	1	DRUG A	10	2026-01-10

input/vs.csv

Edit8 rows

input/vs.csv
STUDYID	VSSEQ	VSTESTCD	VSSTRESN	VSSTRESU
YAMAA-01	1	SYSBP	120	mmHg
YAMAA-01	2	DIABP	80	mmHg
YAMAA-01	3	PULSE	72	beats/min
YAMAA-01	1	SYSBP	ND	mmHg
YAMAA-01	2	DIABP	78	mmHg
YAMAA-01	3	TEMP	98.6	F
YAMAA-01	4	—	—	—
YAMAA-01	1	PULSE	68	beats/min

Expected output

Under review8 expected rows

expected/advs.csv

Edit8 rows

expected/advs.csv
STUDYID	VSSEQ	PARAM	AVAL	LASTDOSE
YAMAA-01	1	Systolic Blood Pressure	120	20
YAMAA-01	2	Diastolic Blood Pressure	80	20
YAMAA-01	3	Pulse Rate	72	20
YAMAA-01	1	Systolic Blood Pressure	—	10
YAMAA-01	2	Diastolic Blood Pressure	78	10
YAMAA-01	3	Other measure	—	10
YAMAA-01	4	Other measure	—	10
YAMAA-01	1	Pulse Rate	68	—

Shaded columns are derived— = missing value

Comments and assessment materials

Dataset Name	Variable Name	Variable Label	Type	Computational Method	Origin
ADVS	STUDYID	Study Identifier	character	Copy value from VS.STUDYID.	Predecessor
ADVS	USUBJID	Unique Subject Identifier	character	Copy value from VS.USUBJID.	Predecessor
ADVS	VSSEQ	Sequence Number	numeric	Copy value from VS.VSSEQ.	Predecessor
ADVS	PARAM	Parameter	character	Recode VS.VSTESTCD: "SYSBP" → "Systolic Blood Pressure", "DIABP" → "Diastolic Blood Pressure", "PULSE" → "Pulse Rate"; missing or unlisted values → "Other measure".	Derived
ADVS	AVAL	Analysis Value	numeric	Copy value from VS.VSSTRESN; a value that cannot be converted → blank.	Derived
ADVS	LASTDOSE	Latest Dose	numeric	Copy value from EX.EXDOSE where EX.EXDOSE IS NOT NULL, keeping the last ordered by EX.EXSTDTC; an absent source → blank.	Predecessor

Version	Date	Author	Description
1.3	2026-09-29	yamaa docs build	Show study codelist names, external dictionaries, and ISO 8601 formats in variable rows; list the document's codelist values and metadata
1.2	2026-09-29	yamaa docs build	Classify ADaM copies from a declared input as Predecessor and recodes as Derived, and SDTM recodes of locally derived columns as Derived
1.1	2026-09-29	yamaa docs build	Resolve lookups into the donor record they select, and give every derivation shape a plain-language rule instead of its raw YAML
1.0		yamaa docs build	Initial generation from spec.yaml

run.R

Edit4 lines

1library(yamaa)23advs <- yamaa_domain("spec.yaml")$output4advs

run.py

Edit4 lines

1import yamaa23advs = yamaa.yamaa_domain("spec.yaml").output4advs

input/ex.csv

input/vs.csv

expected/advs.csv

Comments

Mapping spec

yamaa code

run.R

run.py