AmaraBell0687 · 2 Sept 2025, 19:49 UTC
Separate the coupon identity from each inspection attempt. Count items with an authorised accepted disposition, not rows saying pass. Keep the attempts linked so the report can explain how each current decision was reached.
15 pointsWe have coupon labels but reuse them between batches. Attempts have their own sequence numbers.
3 pointsThen include batch context in the item identity. Also check when those attempt numbers restart. Otherwise a correction to the count may accidentally merge different batches instead of only removing duplicate deliveries.
23 pointsCan a coupon return to review after it was accepted?
9 pointsYes. Quality can reopen it. Current report never removes its old pass from the total.
7 pointsThen quality needs to define the current disposition rules, including reopened work, before anyone chooses a clever query to count them.
16 pointsAsk for that rule in examples: failed then approved retry, repeated delivery of one result, and accepted then reopened. They are different histories even if the current report adds them all happily.
20 pointsQuality supplied examples. Reopened means pending review, outside accepted total. Previous decisions stay in history.
17 pointsTry an out-of-order result too. Newest arrival needn't be newest inspection.
5 pointsYes, and use the documented attempt and review relationships to apply dispositions, not arrival time alone. That avoids an old delayed result changing the current item decision by accident.
12 pointsDeveloper has that case. Duplicate deliveries now leave the offline total unchanged. Reopened coupons disappear from accepted.
7 pointsWhere do results with an unreadable coupon label go?
12 pointsSeparate unmatched queue. They aren't assigned to a coupon by guessing the nearest time.
19 pointsCan somebody see that queue without accidentally accepting its contents?
3 pointsKeep viewing separate from disposition actions. Also check that replaying the saved history rebuilds the same current item list, not only the same total; two different mistakes can cancel in a number.
22 pointsHas the responsible inspector compared that rebuilt list with the batch disposition records yet?
12 pointsYes. Lists match after the replay tests. Inspector approved the report; unmatched remains visible and separate.
16 pointsThanks, Thomas. The delayed-result example exposed a mistake our duplicate test never touched.
19 pointsRetain those examples with the report checks. The next developer needs to know why a late pass is not automatically permission to change the current disposition.
17 pointsThis discussion is closed to new replies after six months without activity. Last activity: .