Predicates - yamaa

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 standards: 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: show which branch of a first-match case a record takes when an input is missing: a comparison with a missing value is unknown rather than false, and only a true condition selects a branch.

Input: one spec.yaml over DM, carrying each subject's age, sex, and treatment group code. Age and group may be missing, and a code may be written in lower case.

Columns:

AGERISK is HIGH over 65, MEDIUM from 40 through 65, and LOW otherwise. The first true branch wins, so an age over 65 is HIGH although it also passes the test for 40. A missing age is unknown in both tests, never true, so it falls through to LOW.
SENIOR_M is Y for a male younger than 65 (its age test is negated), U when age is missing, and N otherwise. Negating an unknown comparison leaves it unknown, so a missing age never reaches Y and is caught by the explicit missing-age test instead.
TRTCLASS copies a screening code (one starting SCRN) through unchanged, and is DRUG_FAMILY for a code starting DRUG, STANDARD for PLACEBO, MISSING when the code is missing, and OTHER for any other code. Pattern matching is case-sensitive, so a lower-case drug-low is OTHER. A missing code is unknown, not true, in the test for "neither DRUG nor PLACEBO", so it reaches MISSING.

Note: a branch is taken only when its condition is true. A comparison with a missing value is unknown, and so is its negation, so a missing value never satisfies a "not" test either; only an explicit test for missingness catches it.

YAML specification

1 spec file

spec.yaml

Edit71 lines

1schema_version: "1.0"2domain: ADSL3keys: [USUBJID]4input:5  DM: {path: input/dm.csv, types: {AGE: int}}6base: DM78output:9  path: adsl.csv10  columns: [STUDYID, USUBJID, AGERISK, SENIOR_M, TRTCLASS]1112columns:13  - name: STUDYID14    type: str15    label: Study Identifier16    derivation: {literal: YAMAA-01}1718  - name: USUBJID19    type: str20    label: Unique Subject Identifier21    derivation: DM.USUBJID2223  # The first TRUE branch wins: subject 01 matches both age branches and24  # takes HIGH. A missing age is UNKNOWN in both comparisons, so subject25  # 03 falls through to otherwise. The 40-year-old pins the inclusive26  # >= 40 boundary.27  - name: AGERISK28    type: str29    label: Age Risk Category30    derivation:31      case:32        - when: "DM.AGE > 65"33          then: {literal: HIGH}34        - when: "DM.AGE >= 40"35          then: {literal: MEDIUM}36        - otherwise: {literal: LOW}3738  # Y marks a male under 65. NOT UNKNOWN is UNKNOWN, not TRUE: subject39  # 03's missing age skips the first branch and is caught by the explicit40  # IS NULL test, while the subjects whose predicate is FALSE fall41  # through to otherwise.42  - name: SENIOR_M43    type: str44    label: Male Under 65 Flag45    derivation:46      case:47        - when: "NOT (DM.AGE >= 65) AND DM.SEX = 'M'"48          then: {literal: Y}49        - when: "DM.AGE IS NULL"50          then: {literal: U}51        - otherwise: {literal: N}5253  # LIKE is case-sensitive: 'drug-low' matches neither LIKE branch, while54  # 'DRUGX' matches 'DRUG%'. NOT IN over a missing value is UNKNOWN, not55  # TRUE, so subject 04 skips to the IS NULL branch. The first branch56  # uses the bare-source shorthand, copying the matched value through.57  # DRUG is caught by 'DRUG%', so only PLACEBO reaches otherwise.58  - name: TRTCLASS59    type: str60    label: Treatment Class61    derivation:62      case:63        - when: "DM.TRTGRP LIKE 'SCRN%'"64          then: DM.TRTGRP65        - when: "DM.TRTGRP LIKE 'DRUG%'"66          then: {literal: DRUG_FAMILY}67        - when: "DM.TRTGRP NOT IN ('DRUG', 'PLACEBO')"68          then: {literal: OTHER}69        - when: "DM.TRTGRP IS NULL"70          then: {literal: MISSING}71        - otherwise: {literal: STANDARD}

Inputs

Under review1 source file

input/dm.csv

Edit6 rows

input/dm.csv
AGE	SEX	TRTGRP
70	M	DRUG
40	F	PLACEBO
—	M	drug-low
25	F	—
18	M	DRUGX
55	F	SCRN-01

Expected output

Under review6 expected rows

expected/adsl.csv

Edit6 rows

expected/adsl.csv
STUDYID	AGERISK	SENIOR_M	TRTCLASS
YAMAA-01	HIGH	N	DRUG_FAMILY
YAMAA-01	MEDIUM	N	STANDARD
YAMAA-01	LOW	U	OTHER
YAMAA-01	LOW	N	MISSING
YAMAA-01	LOW	Y	DRUG_FAMILY
YAMAA-01	MEDIUM	N	SCRN-01

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	Set constant value "YAMAA-01".	Assigned
ADSL	USUBJID	Unique Subject Identifier	character	Copy value from DM.USUBJID.	Predecessor
ADSL	AGERISK	Age Risk Category	character	If DM.AGE > 65 then "HIGH"; If DM.AGE >= 40 then "MEDIUM"; otherwise "LOW".	Derived
ADSL	SENIOR_M	Male Under 65 Flag	character	If NOT (DM.AGE >= 65) AND DM.SEX = 'M' then "Y"; If DM.AGE IS NULL then "U"; otherwise "N".	Derived
ADSL	TRTCLASS	Treatment Class	character	If DM.TRTGRP LIKE 'SCRN%' then "DM.TRTGRP"; If DM.TRTGRP LIKE 'DRUG%' then "DRUG_FAMILY"; If DM.TRTGRP NOT IN ('DRUG', 'PLACEBO') then "OTHER"; If DM.TRTGRP IS NULL then "MISSING"; otherwise "STANDARD".	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

run.R

Edit4 lines

1library(yamaa)23adsl <- yamaa_domain("spec.yaml")$output4adsl

run.py

Edit4 lines

1import yamaa23adsl = yamaa.yamaa_domain("spec.yaml").output4adsl

input/dm.csv

expected/adsl.csv

Comments

Mapping spec

yamaa code

run.R

run.py