TOPIC
Data modeling
These are the same questions found in the library, grouped here by the capability each one mainly examines.
138 questions
Questions 101–138 of 138 on this page.
A · Professional problem solving and implementation3
- The business changes “churn” from 60 days without login to 30 days without payment, while historical training and models use the old definition. How would you manage versions and comparisons?
- A user starts a large export and disconnects; access is revoked while the job runs. The job needs resumability without leaking data after revocation. How would you design it?
- A live dashboard uses event-time windows, while finance groups by settlement time and excludes unsettled transactions. A leader believes one report must be wrong. How would you explain and govern this?
No questions in this section match your target. ·
B · Business understanding and domain modeling25
- An online feature service can return its last computed value after a cache expires, but some predictions need the freshest feature while others tolerate brief staleness. How would you define freshness metadata, task constraints, and unavailable-value fallback?
- A new workflow requires every record to have an ownership field, but historical data lacks it and older clients still write records. How would you design backfill provenance, dual-read/write migration, validation gates, and retirement of old clients?
- An employee creates a project space and later hands it to another owner; the department manager still approves its budget, while audit requires the original creator to remain visible. How would you separate resource ownership, operational responsibility, authorization, and historical attribution?
- A user requests personal-data deletion, but the system uses append-only events, search indexes, and offline analytics copies; some events must remain for business audit. How would you distinguish deletable fields, identity mappings, cryptographic erasure, and retention exceptions?
- A prediction formerly used individual history, but a product now recommends at the organization level; membership changes over time and a person may belong to several organizations. How would you define feature ownership, membership validity, and aggregation deduplication?
- After an upstream migration, a feature keeps falling even though business activity is stable; consumers did not know the source field had been deprecated. How would you define feature lineage, health state, and replacement-source cutover?
- The system changes its user key from email to an immutable ID, but mobile caches, partner callbacks, and historical exports still reference email. How would you design mappings, a dual-read period, privacy cleanup, and retirement of old references?
- The team wants to move eligibility checks from service code to a configurable rules engine so Operations can adjust them; many historical exceptions exist and execution order is implicit in code. How would you discover rules, verify semantic equivalence, and govern configuration access?
- An early consumer deduplicates only by order ID, but one order can have several legitimate changes; the key must include an event version and historical messages need replay. How would you migrate keys without losing or duplicating processing?
- A customer entity must be split into two real customers after an erroneous merge, while old orders, permissions, and metrics reference the original ID. How would you model the split, historical attribution, current-reference migration, and indeterminate records?
- Several teams define user-status events differently, and the platform plans to unify them; some consumers rely on special meanings in old fields and cannot migrate at once. How would you build compatibility adapters, parallel validation, and deprecation governance?
- A product changes permanently valid credits to monthly refreshes, but some older contracts promise permanent balances and users have accumulated unused credits. How would you version rules, balance lots, contract cutovers, and user explanations?
- A new model requires additional features and changes the meaning of existing ones, while online serving must still support rollback to the old model. How would you version feature contracts, deployment order, dual-read cost, and retirement criteria?
- Revenue was recognized for an order last month, and a partial refund occurs this month. Finance and growth disagree on order count, net revenue, and refund-month treatment. How would you establish consistent data semantics?
- A service rebuilds state from an event log, but the live snapshot differs and responders cannot tell which to repair. How would you define authority and the repair process?
- A promotion starts at local midnight, services run in UTC, and customers may move regions. How would you represent effective time and handle boundary invoices?
- Last month’s usage arrives this month. The contract bills monthly and paid invoices cannot be silently overwritten. How would you handle adjustments and notification?
- Payment is authorized but not captured, and goods are not fulfilled. The business report counts an order, while finance records no revenue yet. How would you design cross-domain metric semantics?
- Risk teams need to reproduce decision inputs from three months ago, but the online feature store keeps only the latest state to save cost. How would you balance reproducibility and storage cost?
- A payment channel charges a fee on partial refunds, and the contract assigns it to the merchant, but reports combine refund amount and fee. How would you model this?
- A user belongs to several organizations. The browser retains prior-workspace cache, and old organization data briefly appears after switching. How would you design the boundary?
- A plan has a free allowance, tiered unit prices, and a monthly cap. An upgrade changes the cap immediately, while the allowance resets monthly. How would you define calculation order and explain the invoice?
- Operations counts revenue when an order is placed, while finance recognizes it after fulfillment and refunds. A leader wants one number. How would you explain the two measurement methods and design shared measures for decisions?
- A customer changes plans mid-year. The contract permits partial rollover but forbids stacking discounts. How would you define rollover, reversal, renewal, and invoice transition rules?
- Several teams manually corrected an aggregate table, and later no one could explain why the number changed. How would you record sources, transformations, versions, owners, and recomputation evidence?
No questions in this section match your target. ·
C · Value judgment and direction setting3
- A new operations dashboard is visited often, but interviews show most people only take screenshots before reviews and do not adjust resources or actions. How would you decide whether it creates decision value or only serves a display need?
- The team changes “active customer” from any login to completing a key task and updates the dashboard midquarter; leadership still compares year-over-year trends across the old and new definitions. How would you show the change and limits of comparability?
- A product change increases setup completion, but renewal and long-term use take months to observe; the team wants to invest more based on the proxy. How would you assess its link to final outcomes and the risk of deciding now?
No questions in this section match your target. ·
D · Delivery and ongoing operations5
- A consumer failure leaves a queue hours behind; after repair, many old messages will replay into billing, notifications, and analytics, which have different capacities and side effects. How would you control recovery rate and protect consumers?
- A new system must gradually replace a legacy one while both serve real users, with data and functionality switching in batches. How would you decompose the migration, validation points, and rollback conditions?
- Customer escalations have risen sharply over the past month. Frontline staff blame a few complex customers, while leaders suspect a broken process. How would you distinguish isolated cases from a systemic issue and validate it?
- When channel reports conflict, how would you find differences in their counting rules, time windows, and data sources?
- Upstream events are delayed by hours, while downstream reports still show success and customers make incorrect decisions. How would you design freshness states and notifications?
No questions in this section match your target. ·
H · Professional responsibility and work style2
- A data platform has no direct revenue, but several teams use it to improve analysis and operations. How would you describe platform contribution and its limits?
- A partner plans to describe an unverified analysis in public materials as a “customer-validated result” and asks you to confirm the technical wording. How would you respond?
No questions in this section match your target. ·