Reject Repeated End-of-Study Records

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: Is this a specification error a study team could realistically make, and should the build stop on it rather than continue?
Inputs: Do they trigger the failure for the reason the summary gives?
Expected failure: Is this the right error, raised at the right point, and would it tell the programmer what to fix?

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

Summary

EditUnder reviewWhat this benchmark means

README.md

Goal: derive each subject's end-of-study date (EOSDT) from the subject's one end-of-study disposition record.

Input: demographics records with disposition records carrying category (DSCAT), decoded term (DSDECOD), and start date (DSSTDTC).

Variables:

EOSDT would be the start date of the subject's end-of-study record, the disposition event that is not a screen failure, and blank when the subject has none. The study expects at most one such record per subject; a subject with two has no single end-of-study date, so the run is rejected before any output row is built and no artifact is accepted.

How to fix

Give each subject exactly one eligible end-of-study record in the source data. When the source cannot promise that, drop the uniqueness assertion and choose the record explicitly instead, here the one with the latest start date:

intermediates:
  - id: DS_EOS
    dataset: DS
    filter: "DS.DSCAT = 'DISPOSITION EVENT' AND DS.DSDECOD <> 'SCREEN FAILURE'"
    order_by: [DS.DSSTDTC]
    keep: last
    no_match: null

The no_match line leaves EOSDT blank for a subject with no eligible record, one whose only disposition records are screen failures, instead of failing the lookup.

Keep the uniqueness assertion only for eligible sets that truly carry each key once.

Expected failure

Editthe run is rejected; no artifact is accepted

expected/error.yaml

phase: verification
condition: duplicate_intermediate_records
requirement: REQ-1245

expected/error.yaml

phase: verification
condition: duplicate_intermediate_records
spec_paths:
  - intermediates[0].verifications[0].unique
requirement: REQ-1245
context:
  intermediate: DS_EOS
  dataset: DS
  columns: [STUDYID, USUBJID]
  duplicate_count: 1

YAML specification

1 spec file

spec.yaml

Edit38 lines

1schema_version: "1.0"2domain: ADSL3keys: [STUDYID, USUBJID]4input:5  DM: input/dm.csv6  DS: input/ds.csv7base: DM89output:10  path: adsl.csv11  columns: [STUDYID, USUBJID, EOSDT]1213intermediates:14  # Each subject's end-of-study record: a disposition event other than a15  # screen failure, asserted unique per subject. YAMAA-01-101 carries two16  # of them (a screen-failure record beside them is filtered out and does17  # not count), so the run fails before any output row is built.18  - id: DS_EOS19    dataset: DS20    filter: "DS.DSCAT = 'DISPOSITION EVENT' AND DS.DSDECOD <> 'SCREEN FAILURE'"21    verifications:22      - unique: [STUDYID, USUBJID]2324columns:25  - name: STUDYID26    type: str27    label: Study Identifier28    derivation: DM.STUDYID2930  - name: USUBJID31    type: str32    label: Unique Subject Identifier33    derivation: DM.USUBJID3435  - name: EOSDT36    type: str37    label: End of Study Date38    derivation: DS_EOS.DSSTDTC

Inputs

Under review2 source files

input/dm.csv

Edit3 rows

input/dm.csv
STUDYID	USUBJID
YAMAA-01
YAMAA-01
YAMAA-01

input/ds.csv

Edit5 rows

input/ds.csv
STUDYID	DSCAT	DSDECOD	DSSTDTC
YAMAA-01	DISPOSITION EVENT	SCREEN FAILURE	2024-01-05
YAMAA-01	DISPOSITION EVENT	COMPLETED	2024-01-10
YAMAA-01	DISPOSITION EVENT	WITHDRAWAL BY SUBJECT	2024-01-15
YAMAA-01	DISPOSITION EVENT	COMPLETED	2024-02-11
YAMAA-01	DISPOSITION EVENT	SCREEN FAILURE	2024-03-02

Unexpected Output

Under reviewNo artifact is produced

No fixture files in this directory.

Shaded columns are derived— = missing value

Comments and assessment materials

Dataset Name	Variable Name	Variable Label	Type	Computational Method	Origin
ADSL	STUDYID	Study Identifier	character	Copy value from DM.STUDYID.	Predecessor
ADSL	USUBJID	Unique Subject Identifier	character	Copy value from DM.USUBJID.	Predecessor
ADSL	EOSDT	End of Study Date	character	Take DSSTDTC from the DS record matching on the shared output keys, where DS.DSCAT = 'DISPOSITION EVENT' AND DS.DSDECOD <> 'SCREEN FAILURE', and a row with no match is an error.	Derived

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

How to fix

input/dm.csv

input/ds.csv

Comments

Mapping spec