FA-97931 / Dataset split isolation / Open access

Dataset split: purging uses row positions instead of timestamps · case 01

purging uses row positions instead of timestamps.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The implementation compares each row index to temporal bounds, so its result depends on input order and numeric cutoff values.

THE FAILURE

The implementation compares each row index to temporal bounds, so its result depends on input order and numeric cutoff values.

Unsuccessful approach: Using a strict comparison at the upper bound wrongly purges a record exactly at hi.

Case contract

Given [id,time] rows and validation [lo,hi), retain rows strictly before lo or at/after hi.

Why this case matters

This toy split allocator makes membership rules explicit for reproducible evaluation fixtures.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(rows, lo, hi):
    return sorted(i for j,(i,t) in enumerate(rows) if j<lo or j>=hi)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unordered left right',solve([['c',N+3],['a',N-1],['b',N+1]],N,N+3),['a','c'])
check('exact left boundary',solve([['a',N]],N,N+2),[])
check('exact right boundary',solve([['a',N+2]],N,N+2),['a'])
check('inside',solve([['a',N+1]],N,N+2),[])
check('empty',solve([],N,N+1),[])
check('wide interval',solve([['a',0],['b',N+5]],N,N+5),['a','b'])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
unordered left right	`['c']`	`['a', 'c']`	Failed
exact left boundary	`['a']`	`[]`	Failed
exact right boundary	`['a']`	`['a']`	Passed
inside	`['a']`	`[]`	Failed
empty	`[]`	`[]`	Passed
wide interval	`['a']`	`['a', 'b']`	Failed

SHA-256 / d365d438fee47524da34083e588a68c68218c1da4c1bfbbee351c6ea7f1d0dff

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(rows, lo, hi):
    return sorted(i for i,t in rows if t<lo or t>hi)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unordered left right',solve([['c',N+3],['a',N-1],['b',N+1]],N,N+3),['a','c'])
check('exact left boundary',solve([['a',N]],N,N+2),[])
check('exact right boundary',solve([['a',N+2]],N,N+2),['a'])
check('inside',solve([['a',N+1]],N,N+2),[])
check('empty',solve([],N,N+1),[])
check('wide interval',solve([['a',0],['b',N+5]],N,N+5),['a','b'])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
unordered left right	`['a']`	`['a', 'c']`	Failed
exact left boundary	`[]`	`[]`	Passed
exact right boundary	`[]`	`['a']`	Failed
inside	`[]`	`[]`	Passed
empty	`[]`	`[]`	Passed
wide interval	`['a']`	`['a', 'b']`	Failed

SHA-256 / b3f7e23263008239e68339c3258c148744b4b7823343e6cfe2961db5323e2368

HELD IN THE MEMBER ARCHIVE

The verified repair and its recorded checks are member-only.

This mechanism has 6 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.

Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.

Member access is invitation-based. Sign in with your invited account to inspect the repair.

Verification & scope

Inputs are small in-memory records with self-contained contracts; no claim is made about production sampling quality. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-30T16:48:07.547981+00:00.

Case digest / 7e8408a4f61245e1f0d5cc36b364e726148c379aee398c8c36452dd8703ba281