Benchmark

660 benchmarks, generated from benchmarks/. Each one is a complete, runnable specification with its input data and the exact output an implementation must reproduce, and each page below shows that README, the source fixtures, the expected artifacts, and the YAML specification together.

Group	Count	Purpose
sdtm	70	Assess collected data to SDTM and SEND tabulations
adam	130	Assess SDTM to ADaM and ADaM to results derivations
schema	115	Assess yamaa schema
negative	345	Assess yamaa error handling

Rejection is half the contract. A portable specification language is defined as much by what it refuses as by what it computes, and a refusal is only portable if two implementations refuse the same thing in the same way, so every anti-pattern pins the exact error alongside the specification that provokes it.

The suite serves three audiences at once:

• Users read it as a cookbook: "how do I express a baseline flag" has an answer to copy.
• Developers run it as a conformance test: the expected files are byte-exact contracts.
• AI agents learn the common derivations from one half and the anti-patterns that bound them from the other.

Start here

#	Benchmark	What it establishes
1	Basic Demographics	Reading collected items, handlers, and the declared keys as the grain
2	Lab Findings	Row construction: one template per collected test, row_number for the sequence
3	Lab BDS Build	A full Basic Data Structure build: parameters as row templates, then baseline, change and sequence as columns

Read a page top to bottom and the loop closes in five minutes: the Summary gives the intent, the specification sidebar gives the shape and the grain, and the expected artifact confirms the reading.

All benchmarks

sdtm

70 benchmarks. Each runs to completion and produces the artifact in expected/.

AE

5 benchmarks

• Code Reported Terms Against a Medical Dictionary
• MedDRA Hierarchy from Coder-Assigned LLT Codes
• Repeated Adverse Events from ODM
• Partial Adverse Event Dates
• Seriousness Criteria

BW

1 benchmark

• Tabulate nonclinical body weights with SENDIG 3.1.1

CM

3 benchmarks

• Collected Medications with Dose Text and Timing References
• Repeated Concomitant Medications from ODM
• Code medications against WHODrug

CO

1 benchmark

• Split Long Comments into 200-Character Pieces

DM

11 benchmarks

• Age From Birthdate
• Arm Assignment
• Basic Demographics
• Reference Dates
• Death Date Flag
• Metadata Contract
• Demographics When No Subject Died
• Demographics from an EDC Snapshot Export
• Record race and ethnicity, with one supplemental record per race
• Demographics Submitted in SAS Transport with SDTMIG 3.3
• Demographics Submitted in SAS Transport

DS

3 benchmarks

• Disposition from Several Source Forms
• Discontinuation Reasons
• Disposition Sequence

EC

1 benchmark

• Collected Daily Dosing

EG

1 benchmark

• Replicate ECG Results at Planned Time Points

EX

4 benchmarks

• Combination Regimen
• Exposure from Collected Records
• Infusion duration as ISO 8601 duration text
• Dose Intervals

FA

3 benchmarks

• Event Findings
• Multiple Findings About from ODM
• Reactogenicity Diary

LB

9 benchmarks

• Standardize Non-Numeric Lab Results
• Conditional Compartments
• Build One Record per Collected Lab Result
• CTCAE v5.0 Grading
• Lab-Specific Reference Ranges
• Last Observation Before Exposure
• Value-Level Metadata
• Multiform Labs
• Reference Range Flags

MH

2 benchmarks

• Ongoing Conditions with Partial Dates
• Pre-specified Medical History Checklist

MO

1 benchmark

• Standardize morphology results with SDTMIG 3.3

PC

1 benchmark

• Parse planned times in a concentration transfer

PP

1 benchmark

• Publish parameters from a pharmacokinetic analysis transfer

PR

1 benchmark

• Prior Procedures from Surgery and Radiotherapy Forms

QS

1 benchmark

• Questionnaire Items

RELREC

2 benchmarks

• Dataset-Level Related Records
• Related Records

RS

1 benchmark

• Target Lesion Response at Each Tumor Assessment

SE

1 benchmark

• Subject elements from DM and an ODM extract

SUPPDM

1 benchmark

• Supplemental Qualifier Metadata by Qualifier Name

SUPPMH

3 benchmarks

• Parent Linkage
• Continue Long Medical History Terms Between Words
• Supplemental Qualifiers

SV

2 benchmarks

• Describe actual subject visits with SDTMIG 3.3
• Build SDTM subject visits from ODM extracts

TA

1 benchmark

• Trial Arms with a Sponsor-Added Epoch

TR

1 benchmark

• Tumor Measurements

TS

1 benchmark

• Turn protocol metadata into trial summary parameters

TU

1 benchmark

• Tumor Identification

VS

7 benchmarks

• Collected Vital Signs Form
• Missed Temperature Diary Days
• Assign Each Vital Sign's Trial Period
• Planned Time Points Around Dosing
• Replicate Readings
• Visit Study Day
• Standardize Vital-Signs Units

adam

130 benchmarks. Each runs to completion and produces the artifact in expected/.

ADAE

24 benchmarks

• Analysis Start Time of an Adverse Event with ADaMIG 1.1
• Analysis Start Time of an Adverse Event
• Death Carryforward
• Exposure Adjusted Adverse Event Incidence
• First Event Flags
• Treatment Emergence Timing
• Partial Date Imputation
• Partial Time Imputation
• Treatment emergence by phase in a neoadjuvant, surgery and adjuvant study
• Adverse Events Within 24 Hours After a Dose
• Post-Index Event Flags
• Name Each Event's Prior Serious Event
• Reported Terms in Proper Case
• Protocol Review Flags
• Query Membership Flags
• Review Ordering
• Keep Only Serious Events
• Serious Event Sequence
• Carry Each Event's Severity
• Severity Correction
• Severity Ranking
• Text Cleanup
• Split Adverse Events at Severity Changes
• Worst Severity Flags

ADAECM

1 benchmark

• Pair Each Adverse Event with Each Medication Related to It

ADCE

1 benchmark

• Worst Toxicity Flags

ADCM

5 benchmarks

• ATC Classes for Coded Medications
• Partial End Dates Capped at Death
• Free-Text Start Dates Read as ISO 8601 Dates
• On-Treatment Flags
• List the adverse events each medication was given for

ADEG

6 benchmarks

• Derived ECG Intervals
• Placebo-Corrected ECG Change from Baseline
• Flag QTcF Values and Increases Beyond Categorical Limits
• Replicate ECG Averages with QTcF Corrected per ECG
• Time-Matched ECG Baselines by Period
• Time-Matched Plasma Concentrations for ECG Time Points

ADEX

4 benchmarks

• Cumulative Dose
• Dose Reduction Flag
• Infusion duration from the collected duration or the start and end times
• Uncollected Exposure

ADFA

1 benchmark

• Fever Occurrence

ADICE

1 benchmark

• Governing Intercurrent Event

ADIS

1 benchmark

• Derive log10 neutralizing antibody titers

ADLB

16 benchmarks

• Compare laboratory results with baseline with ADaMIG 1.1
• Lab BDS Build
• Analysis value as fixed-decimal text
• Closest Visit Selection
• Flag the End-of-Treatment Laboratory Record
• Apply the stated policy to qualified laboratory results
• Derive absolute differential records
• Calculate each subject's mean result
• Analysis Value Metadata by Parameter
• Derive sex-specific reference limits from the study's percentiles
• Ordered Float Sum
• Name Each Laboratory Parameter and Visit From the Study's Terminology
• Derive each parameter's reference limits from the study's results
• Normal Limit Check
• Classify a Result, Its Shift from Baseline, and Two Criteria
• Join a Numeric Sequence to a Padded Supplemental Key

ADLBC

2 benchmarks

• Lag a Result Across Visits While Constructing Change Parameters
• Carry a Previous Visit's Change

ADOE

1 benchmark

• Study Eye Flag

ADPC

2 benchmarks

• Set below-limit concentrations by their place in the profile
• Actual Times Since the First Dose and Since the Reference Dose

ADPP

6 benchmarks

• Accumulation Ratios Between Day 1 and Steady State
• Carry-Over Exclusion of a Crossover Period
• Records Each Planned Treatment Comparison Uses
• Add dose-normalized exposure parameters
• Geometric mean metabolite ratio and replicate variability per subject
• Link each PK parameter to the period and treatment of its dose

ADQS

5 benchmarks

• Carry the last post-baseline score to each missed scheduled visit
• Multi-Scale Instrument Scoring
• Phantom Records for Missed PRO Assessments
• Screening Score Used as Baseline
• Subscale Scoring

ADRS

6 benchmarks

• Select the Best Overall Response
• Assign a Composite Responder Status
• Confirmed Response
• Flag Measurable Disease at Baseline
• Prepare Response Assessments for Best Overall Response
• Derive Investigator Overall Response Records

ADSL

25 benchmarks

• Subject-Level Analysis Data Submitted in SAS Transport with ADaMIG 1.1
• Subject-Level Analysis Data Submitted in SAS Transport with ADaMIG 1.2
• Pool Subjects into Age Groups
• Derive the Last Known Alive Date and Death Date from Multiple Domains
• Derive the Analysis Age at Randomization
• Derive BMI Two Ways
• Group subjects by the study's baseline BMI quartiles
• Flag Subjects Who Completed the Study
• Normalize the Collected Country and Assign Its Region
• Derive period-scoped treatments and dates across a washout
• Derive planned period treatments and sequence from the trial design
• Standardize Collected Sex, Race, and Age
• Summarize End-of-Study and End-of-Treatment Disposition
• Flag Any Reported Dose Adjustment
• Derive Exposure Duration in Weeks and Months
• Derive a Chain of Population Flags
• Keep an Investigator Comment Exactly as Collected
• Derive the new anti-cancer therapy start date
• Derive analysis period start and end dates in a two-period study
• Give Each Analysis Period an End-of-Period Status
• Derive Randomization Timing
• First Rescue Medication
• Parse the Site from the Subject Identifier
• Select the Actual Treatment and Its Exposure Dates
• Subject-Level Analysis Data Submitted in SAS Transport

ADTR

2 benchmarks

• Track the Running Nadir Sum
• Sum the Target Lesion Diameters

ADTTE

5 benchmarks

• Duration of Response
• Measure Time to the First Adverse Event
• Overall Survival
• Progression-Free Survival
• Time to Deterioration for a PRO Endpoint

ADVS

14 benchmarks

• Derive BMI
• Derive Body Surface Area
• Once-Measured Carryforward
• Place each vital-signs record in its crossover period
• Weight Used for Dosing
• Carry the First Observed Vital-Sign Result
• Carry the last observed value across missing visits
• Carry one observed record to each planned visit
• Derive mean arterial pressure records
• Derive the Change on Standing and Orthostatic Flags
• Derive growth percentile records
• Retain the Latest Earlier Character Result
• Assign Analysis Windows from the Window Table
• Assign Analysis Windows by Study Day

CLINSITE

1 benchmark

• Clinical Site Data at Site Level

RESULTS

1 benchmark

• Adverse Event Incidence Results

schema

115 benchmarks. Each runs to completion and produces the artifact in expected/.

ADAE

15 benchmarks

• Carry Subject-Level Variables onto Adverse Events by Name
• Date and Time Imputation Flags of an Analysis Start
• Analysis Start Date Kept Within Treatment Start and the Event's End
• Analysis Start Date/Time Falling Back to First Exposure
• Analysis Start Date From a Collected Date and Time
• One Fragment on Two Templates
• Column-Level Values on Rows Built From a Selection of Events
• Give Each Adverse Event the Dose Nearest Its Start
• Carry Several Supplemental Qualifiers Onto Their Adverse Events
• Adverse Events Take the Treatment Period Holding Their Start Date
• Ordering
• Boolean Substring Flags
• Clean Up Collected Text
• A Time of Day as a Column Type
• Window Rank Over Intermediate Records

ADCM

1 benchmark

• Analysis Start Date From Complete Collected Dates Only

ADEG

2 benchmarks

• Select the Concentration Closest in Time to Each ECG
• Give Each ECG the Concentration Nearest in Time

ADEX

5 benchmarks

• Aggregate Bindings Read an Intermediate
• Derive Date Epoch
• Each Dose Level a Subject Received, Against the Planned Dose
• Each Dose Takes the Period Holding the Start Date the Output Carries
• Each Dose Takes the Treatment Period Holding Its Own Start Date

ADLB

9 benchmarks

• Baseline Skips Missing Results
• Baseline Among Same-Day Results
• Column Composition
• Spec Inheritance
• Execution Lifecycle
• Lookup Filters
• Row Templates
• Sentence Case and Title Case
• Show Where Each Column Comes From

ADQS

1 benchmark

• Derive Text Score

ADSL

39 benchmarks

• Keep Out-of-Range Ages for Review
• Aggregate Narrows on Its Derived Record Date
• Count Adverse Events From Each Subject's Own First Dose
• Subject Dose Aggregation
• Birth Date Completion
• Replace One of a Study's Shared Codelists in a Submission Package
• Expression Forms
• Treatment Start and End Datetimes From Dates
• Inherited Entry Removal
• Shared Input, Project Data
• Inherited Input Types
• Inherited Output
• Count Treatment Days From a Saved Demographics File
• Fatal Event Join
• First Dose and Death Dates Stay Blank Without a Record
• Place Each Dose in Its Period to Find the First Dose of Period 2
• Derived Intermediate Rank
• End-of-Study Record
• Report Repeated Exposure Records and Read the Latest
• Take Each Subject's Weight Nearest the Randomization Date
• Study-Level Values Read From a One-Record Settings Table
• Choose the Last Height and the Last Weight of Each Subject
• Derived Demographics Stay Blank When the Collected Value Is
• Non-Finite Values
• Numeric Functions
• Parquet Everywhere
• Predicates
• Share One Study's Project Between Two Submission Packages
• Pin ties half away from zero
• Row Count Bounds
• Short Spellings
• Read a Spreadsheet Export That Starts with a Byte-Order Mark
• Read a Dataset-JSON File
• Read the Times and Weights a Data Frame Library Writes to Parquet
• Read a SAS Transport File
• Text Operations
• Text Mapping
• Uncollected codes stay missing; outside-codelist codes are labelled
• What the Run Read, Built and Checked

ADVS

7 benchmarks

• Baseline Strictly Before First Dose
• Stated Fallbacks
• One Template Replacing a Column Default
• Source Binding
• Read a Newline-Delimited Dataset-JSON File
• Describe Observed and Averaged Values Apart
• Visit-Level Window Functions

AE

2 benchmarks

• Calendar Duration and Interval Text Between Two Dates
• Required Values on Selected Rows and Within a Share

AECM

9 benchmarks

• Count the Medications Related to Each Adverse Event
• Pair Each Adverse Event with Its Medications, Matching Terms in Upper Case
• Pair Each Adverse Event with the Medications Started on or After It
• Count the Medications Given for Each Treated Adverse Event
• Find the Last Medication Related to Each Adverse Event
• Pair Each Adverse Event with the Medications Recorded at Its Visit
• Keep Each Adverse Event's Medication Pairs in Stored Order
• Pair Each Adverse Event with Each Medication of Its Relationship
• Pair Each Adverse Event with the Doses of the Medications Given for It

BANDDOSE

1 benchmark

• Find the first dose ending in the second day band

DI

1 benchmark

• Bind Each Dataset to the Implementation Guide It Names

DM

5 benchmarks

• Block Composition
• Submit demographics with the older metadata format
• Replace Collected Sex with a Reviewed Unknown Value
• One Collected Answer per Subject
• Newline-Delimited Dataset-JSON

IDS

1 benchmark

• Read a One-Column List That Ends with a Blank Line

LB

5 benchmarks

• Describe a Result Reported by Two Laboratories
• Form-Scoped Items
• Lab Results with Their Own Sample Dates
• Describe Central and Site Laboratory Results Apart
• Numbered Results Each Keep Their Own Values

LESION

1 benchmark

• Impute a Missing Outcome from the Nearest Donor

OUT

2 benchmarks

• Select a Measurement Closest to Each Reference Day
• Window Home Measurements by Their Own Study Day

RANGES

2 benchmarks

• Adjacent Windows Share an Endpoint
• Window Endpoints Decide Which Records Match

RECORDS

1 benchmark

• Source record ordinal

TEST

2 benchmarks

• Call a Project Routine
• Preserve concentration units using the received text encoding

TIMELINE

1 benchmark

• Start Dates Read From Each Row's Own Record

VS

2 benchmarks

• A Column Default Reads Its Own Driver Beside Another Driver's Groups
• Numbered Measurements Each Read Their Own Item Group

ZE

1 benchmark

• Submit sponsor-defined examination records

negative

345 benchmarks. Each must fail, and expected/error.yaml pins the error it raises.

ADAE

34 benchmarks

• Reject taking the subject identifier from the subject-level dataset again
• Reject a subject-level variable the subject-level dataset does not carry
• Reject subject-level variables read from a file without its producer
• Reject Date Extraction
• Reject Incomplete Date
• Reject Mixed Zones
• Reject an Imputation Flag Read From Collected Text
• Reject Nonexistent Day
• Reject Impossible Month
• Reject Non-Date Source
• Reject Unknown Day
• Reject an Onset Recorded as a Period
• Reject an Adverse Event Whose Subject Has No Subject-Level Record
• Reject Source Domains Checked Without the Study's Dataset List
• Reject a Saved Subject-Level Dataset Older Than Its Description
• Reject an Event With Two Qualifier Records of One Name
• Reject a Qualifier Record Whose Name Is Not Declared
• Reject Repeated Order
• Reject Non-Date Flag
• Reject Slot Overflow
• Reject Invalid Ranking
• Reject Date Arithmetic in a Condition
• Reject Date-Text Comparison
• Reject Unknown Date
• Reject a Record Selection Stated Twice
• Reject Word Separators on Title Case
• Reject Invalid Regex in str_contains
• Reject Invalid Regex in str_replace
• Reject a Study Day Stored as a Date
• Reject Non-Numeric Sum
• Reject a Time of Day Taken from Text
• Reject a Time Taken from Text
• Reject an Empty Adverse Event Dataset in SAS Transport
• Reject an Event Numbering That Groups by the Subject Twice

ADCM

2 benchmarks

• Reject Two Day and Month Orders for Slashed Start Dates
• Reject a Start Date in an Unlisted Spelling

ADEG

2 benchmarks

• Reject Rederived RR
• Reject Two Ways of Choosing the Sample for an ECG

ADEX

9 benchmarks

• Reject Uncomputed Dose
• Reject Dose Expansion
• Reject Per-Record RDI
• Reject a Dose Group Written as a Number and as Text
• Reject a Day Count from Text
• Reject Unit Suffix
• Reject a Planned Dose Found While the Records Are Still Being Built
• Reject a Value Stated as Both a Total and a Single Dose
• Reject Unknown Field

ADIS

1 benchmark

• Reject the base-10 logarithm of a zero titer

ADLB

37 benchmarks

• Reject Self-Referential Parameter
• Reject Duplicate WBC
• Reject Duplicate Baselines
• Reject Ambiguous Type
• Reject Baseline Tied on Date and Time
• Fail an Unnamed Count on Bad Data
• Reject a Parameter Decoded Through a Format Name
• Reject a Visit Name Read From Text Codes by a Visit Number
• Reject a Parameter Name Read From an Undeclared Terminology
• Reject a Parameter Code the Study's Terminology Does Not List
• Reject a Parameter Name From a Terminology That Lists No Names
• Reject Parameter Names Read Without the Study's Terminology
• Reject Formula Flag
• Reject Forward Reference
• Reject Integer Overflow
• Reject Log Zero
• Reject an Analysis Window Placed by a Day the Dataset Does Not Carry
• Reject Duplicate Key
• Reject Partial Key
• Reject Unmapped Key
• Reject Multiple Baselines
• Reject Below-Limit Value
• Reject Folder Path
• Reject a File Path With an Empty Folder Name
• Reject Missing Table
• Reject Symlink Path
• Reject URL Path
• Reject a quantile beyond the greatest value
• Reject Row Aggregate
• Reject Numeric Filter
• Reject Self Reference
• Reject fixed-decimal text made from a text result
• Reject a zero text width
• Reject a Nonnumeric Padding Width
• Reject Tied Baseline
• Reject a Baseline Check That Matches Subjects Through Another Dataset
• Reject Zero Division

ADQS

1 benchmark

• Reject a Device Questionnaire Placed by a Clinic Date

ADRS

2 benchmarks

• Reject Unmapped Value
• Reject Post-CR PR

ADSL

123 benchmarks

• Reject an Analysis Age Described as Collected From the Subject
• Reject a Tabulation Core Designation on an Analysis Column
• Reject an Analysis Date Submitted as Text
• Reject a Length Declared on an Analysis Date
• Reject an Event Window Opened by Another Dataset's Date
• Reject One Summary Over Exposure and Vital Signs Records Together
• Reject a Total That Sums an Unnamed Per-Record Value
• Reject Two Inputs Without a Base Dataset
• Reject Flag With Only a Fallback Value
• Reject a Package Codelist That Reuses the Study's Name Under Another Code
• Reject a Numeric Age Group Bound to a Text Codelist
• Reject a Copied Age That Also Claims to Be Derived
• Reject Two Datasets That Each Take a Value From the Other
• Reject Coded Age
• Reject a Boolean Dataset-JSON Column
• Reject a Dataset-JSON File Cut Off in Transfer
• Reject a Dataset-JSON File That Does Not Count Its Records
• Reject a Dataset-JSON Record With a Value Missing
• Reject a Dataset-JSON Value Against Its Column's Type
• Reject a Dataset-JSON 1.0 File
• Reject Mixed Endpoints
• Reject Mixed Units
• Refuse a Data Definition That Points Outside Its Package
• Refuse a Copied Age Labelled Differently From Its Source
• Reject a Derived Age Group Without Its Method
• Reject a Baseline Body Mass Index From an Unchosen Screening Weight
• Reject Broken Flag Condition
• Reject Flag Without Missing Value
• Reject Numeric Flag
• Reject a Dose per Kilogram From a Zero Weight
• Reject a Call to a Project Routine the Project Does Not List
• Reject a geometric mean over a zero ratio
• Reject Mixed Scales
• Reject a Randomization Matched on One Administration's Site
• Reject a Treatment Period Found From One Administration's Day
• Reject Implausible Age
• Reject a Date and Time Written With a Space
• Reject Published-Variable Metadata on a Working Value
• Reject Invalid Sex
• Refuse arithmetic over a list of sequence numbers
• Refuse an adverse event list in no stated order
• Reject a Subject Record Identified by the Subject's Highest Dose
• Reject Conflicting Ages
• Reject a Site-Scoped Subject Identity
• Reject Mixed Types
• Reject a Saved Dataset Typed by the File That Reads It
• Reject Structured Literal
• Reject Dose Periods Placed by the Subject's First Dose Date
• Reject Repeated End-of-Study Records
• Reject a Last Record Chosen from No Order
• Reject Self-Named Treatment
• Reject a Matching Column Listed Twice
• Reject Unchosen Record
• Reject a Repeated-Record Warning With No Record to Read
• Reject Unordered Choice
• Reject a Subject Without a Planned Treatment
• Reject Case Collision
• Reject Unreadable Pattern
• Refuse a quantile beside a mean
• Reject Negative Sqrt
• Reject a Largest Daily Dose Written as One Summary Inside Another
• Reject Missing Age
• Reject Too Many Missing Baseline Weights
• Reject a Missing-Weight Limit Above One
• Reject an Age Group Format Name With the Dollar Sign
• Reject Duplicate Subject
• Reject Transport Container Output
• Reject Dates That Are Not Paired
• Reject an Empty Name for the Shared File
• Reject Shared Definitions That Are Not There
• Reject a Body Surface Area From a Negative Weight
• Reject Datetime Precision of a Zoned Date and Time
• Reject a Copied Age That Does Not Name Its Source
• Reject a Study Project That Inherits From Its Own Package
• Reject a Package Whose Study Project Is Not There
• Reject a Package That Names Two Parent Projects
• Reject Cleared Property
• Refuse a quantile without its proportion
• Reject an Age Check Whose Limits Are Reversed
• Reject Redefined Date
• Reject Remote Parent
• Reject a Run Report Saved Over Its Dataset
• Reject rounding a coded value
• Reject an Order of First Doses Taken From Single Administrations
• Reject an Analysis Dataset Listed With the Tabulation Datasets
• Reject Bad Sex
• Reject a Lone Check Not Written as a List
• Reject Bad Encoding
• Reject a True-or-False Field in a Parquet Export
• Reject a Byte-Order Mark on a Parquet Input
• Reject a Carriage Return Inside a Field
• Reject Duplicate Key
• Reject an Empty Export File
• Reject Extra Field
• Reject Duplicate Field
• Reject a Randomization Time With a Fraction of a Second
• Reject a Missing Byte-Order Mark
• Reject NA Age
• Reject a Quotation Mark in an Unquoted Comment
• Reject a Subject Without a Dosed Exposure Record
• Reject Text After a Closing Quote
• Reject an Undeclared Byte-Order Mark
• Reject Undeclared Field
• Reject Unknown Format
• Reject Unnamed Field
• Reject Unterminated Quote
• Reject Numeric Lowercase
• Reject Unknown Case
• Reject a Site Entered Without Its Site Number
• Reject Uncaptured Group
• Reject Strata Mismatch
• Reject Study Datetime
• Reject Malformed Reference
• Refuse a total of distinct doses
• Refuse a total that states an order
• Reject a Birth Date Before 1582 in SAS Transport
• Display Digits Written into Submitted Heights
• Reject a Uniqueness Check Over Events Computed Per Subject
• Reject a Uniqueness Check Over Events Chosen Per Subject
• Reject Duplicate SITEID
• Reject a Second Statement of the Row Identity
• Reject Nested Expression
• Reject a Parent Written for a Newer Language Version

ADVS

20 benchmarks

• Reject Fractional Pulse
• Reject Self Fallback
• Reject a Baseline Flag Listed Before Its Visit Number
• Reject a Key That Depends on a Later Column
• Reject a Change From Baseline That Reads an Undeclared Value
• Reject carry-forward without an assessment order
• Reject a Period Found While the Records Are Still Being Built
• Reject a z-score for a percentile of 100
• Reject a Missing Output Key
• Reject Overlapping Windows
• Reject Missing Prior
• Reject a Baseline Flag Set Before Visits Are Numbered
• Reject Future-Filled Weight
• Reject a standard deviation of text results
• Reject an undefined visit ordering
• Reject a Record Two Value Descriptions Select
• Reject a Comparison of One Value Written for Several
• Reject a Systolic Record Without a Result
• Reject Two Value Descriptions That Can Describe One Record
• Reject a Value Description Selected by a Variable the Dataset Lacks

AE

5 benchmarks

• Reject a calendar duration from a start date held as text
• Reject a Coding Dictionary That Is Also a List of Terms
• Reject a Collected Month Written as One Digit
• Reject a Reported Event Without Its Coded Term
• Reject Text Read as a Whole Number from a SAS Transport File

AECM

5 benchmarks

• Reject Two Treating Medications Where Each Adverse Event Has One
• Reject a Choice of One Medication Beside the Pairing
• Reject a Rescue Medication Named by a Word Where a Number Is Recorded
• Reject Pairing Events with Medications Without Saying What They Share
• Reject Pairing Adverse Events with the Rows Being Built

CM

1 benchmark

• Reject Rename-Only Intermediate

CO

2 benchmarks

• Reject the Length of a Comment Number
• Reject a Comment Piece Counted from Zero

DI

1 benchmark

• Reject a Device Dataset Whose Implementation Guide the Study Omits

DM

58 benchmarks

• Reject a Second List of Sex Codes
• Reject a Sponsor Addition to a Closed Published Codelist
• Reject a Codelist That Decodes Only Some of Its Values
• Reject a Binding to a Codelist the Project Does Not Declare
• Reject a Codelist That Lists a Value Twice
• Reject Sex Codes Checked Without the Study's Terminology
• Reject a Computed Age Described as Collected
• Reject Empty Text Kept as a Value in a Comma-Separated File
• Reject a Partial Birth Date Submitted as a Complete Date
• Refuse a Data Definition Without the Annotated Case Report Form
• Refuse a Data Definition of a Variable Without Its Metadata
• Refuse a Data Definition That Points to an Undeclared Document
• Refuse a SAS Transport File With an Unreadable Display Format
• Reject a Sex Format Name Without the Dollar Sign
• Reject a Subject Record Identified by the Subject's Collected Sex
• Reject a Subject Identifier With Two Different Length Limits
• Reject a Site Identifier With a Length of Zero
• Reject a Dataset Label Kept in Free-Form Notes
• Reject an Item Read for Rows Built Elsewhere
• Reject a Subject Without a Subject Key
• Reject an Item Defined Twice With Two Names
• Reject an ODM Export Declared in Latin-1
• Reject an ODM Export Without Collected Data
• Reject an Answer Recorded at Two Visits
• Reject an Unknown Null Marker in an ODM Export
• Reject an Item That Is Null and Answered
• Reject a Record the Extract Repeats
• Reject Two Fields That Name One Value
• Reject an Extract Without Its Form Level
• Reject Declared Types on ODM Data
• Reject a Record Without Its Subject
• Reject a Transactional ODM Export
• Reject Typed Item Elements in an ODM Export
• Reject an Unescaped Ampersand in an ODM Export
• Reject a Study Subject Number Given by Two Vendors
• Reject an ODM 1.2 Export
• Reject a Submitted Variable That States No Origin
• Reject a Demographics Record Without Sex
• Reject a Role the SDTM Model Does Not Publish
• Reject an Analysis Guide Cited for the Tabulation Datasets
• Reject Two Versions of One Implementation Guide in One Data Definition
• Reject Duplicate Sex
• Reject Missing Field
• Reject Trivial Filter
• Refuse to Submit Demographics Last Saved Under an Earlier Name
• Reject a Subject Identifier Built Without Its Site
• Reject Submission Metadata That Names No Model
• Reject a Terminology Standard Define-XML Does Not Name
• Reject Two Releases of SDTM Terminology in One Data Definition
• Reject a Country Submitted Without a Length
• Transport Beside a Document Not Named define.xml
• Empty Investigator Name in Transport
• Sex Label Too Long for Transport
• Ethnicity Name Too Long for Transport
• Accented Investigator Name in Transport
• Site Identifier with a Trailing Blank in Transport
• Reject Two Text Operations Written as One Value
• Reject Long Identifier

EX

1 benchmark

• Reject a duration written from minutes held as text

LB

11 benchmarks

• Reject a Reference Range Matched on a Looked-Up Sex
• Reject Mismatched Lists
• Reject Missing Sex
• Reject an Age Band Chosen by Another Dataset's Age
• Reject a Reference Limit That Is Both Required and Optional
• Reject a Lookup That Names No Match Columns
• Reject Unmatched Key
• Reject Per-Test Lengths Over a Numeric Test Code
• Reject a Laboratory Result Without a Unit Conversion
• Reject Two Descriptions That Each Exclude the Other's Tests
• Reject a Converted Laboratory Result Described as Delivered

LESION

1 benchmark

• Refuse a Reduction Inside a Donor Selection

QS

2 benchmarks

• Reject a Dataset Name With a Hyphen
• Reject a Diary Selection Checked Against Another Dataset's Date

RECORDS

1 benchmark

• Reject a source ordinal name collision

SUPPDM

1 benchmark

• Reject a subject-wide count inside a one-race row

SV

1 benchmark

• Reject Missed Visits Chosen Before Any Visit Is Recorded

TA

1 benchmark

• Reject a Sponsor-Added Epoch That Claims a Published Code

TE

1 benchmark

• Reject Reference Data Declared as Repeating

TEST

9 benchmarks

• Reject Cyclic Parent
• Reject a Missing Required Argument
• Reject Two Project Routines With One Name
• Reject a Non-Finite Result From a Study Calculation
• Reject a Project Routine Whose Own Test Case Fails
• Reject a Template That Adds the Study File
• Reject a Fragment That Names Its Own Template
• Reject One Fragment Added Twice
• Reject a text encoding that contradicts the received file

VS

12 benchmarks

• Reject Device Calibrations Matched on a Number the Dataset Does Not Carry
• Reject a Calibration Count Matched on a Record Some Rows Do Not Have
• Reject an Epoch the Study Never Declared
• Reject Incomparable Endpoints
• Reject a Visit Window Found From the Clinic Record for Home Rows
• Reject a Device Matched on a Record Some Rows Do Not Have
• Reject Unknown Column
• Reject a Daily Home Blood Pressure Taken From One Reading
• Reject a Vital Signs Date-Time That Would Invent Seconds
• Reject a First Visit Window Taken From Records Not Yet Built
• Reject a Value Description Selected by a Test Code the Study Never Lists
• Reject a Sequence Number Ordered by the Clinic Visit Day

ZE

1 benchmark

• Reject an older submission format for sponsor-defined records