← 返回 oracle 的题目列表OOD — Treatment Frequency Scheduler
类型：qbank
Design classes that schedule patient treatments at varying frequencies within a date range, then enumerate every required treatment instance. Asked as the OHAI phone-screen problem with explicit allowance to write pseudo-code or comments for date arithmetic syntax.
Requirements
Model a treatment with a configurable repeat frequency (e.g. every 4 hours, every 12 hours, daily at 9 AM).
Given a treatment and a date range [start, end], produce the list of all treatment times that fall within the range.
Multiple treatments per patient; the output should be sortable and groupable per patient.
Date arithmetic syntax is not graded — the interviewer explicitly allowed leaving the date math as comments. Logical structure of the classes is what's scored.
Notes
Core class skeleton:
enum FrequencyUnit { MINUTES, HOURS, DAYS, WEEKS }
class TreatmentSchedule {
String name;
int every; // every N units
FrequencyUnit unit;
Instant anchor; // first scheduled time
// Optional: time-of-day constraint, weekday filter, etc.
List<Instant> occurrencesWithin(Instant from, Instant to) { ... }
}
class PatientTreatmentPlan {
String patientId;
List<TreatmentSchedule> schedules;
List<(String, Instant)> allOccurrences(Instant from, Instant to) { ... }
}
The occurrencesWithin method is the core algorithm: starting from anchor, generate anchor + k * frequency for increasing k while the result falls within [from, to]. Skip generated times before from.
If a treatment has a time-of-day constraint (e.g. "every day at 9 AM"), encode it as (anchor: 9:00 AM today, every: 1, unit: DAYS) rather than a special-case field.
Trap: do not eagerly precompute all occurrences across the patient's lifetime. The query is range-bounded; iterate lazily within [from, to].
Trap: time zones. The interviewer didn't ask, but in healthcare scenarios time-zone-aware scheduling matters — call this out explicitly.
Don't over-engineer with strategy patterns / factories before requirements emerge — same lesson as the body-temperature variant of this prompt.
Preparation
Draft the two-class skeleton in 15 minutes.
Implement occurrencesWithin as a tight loop; verify with a 4-hour treatment over a 24-hour range produces exactly 6 entries.
Have the time-zone caveat ready as a follow-up comment, even if the prompt doesn't surface it explicitly.