How should I practice In-Memory Database with TTL and Historical Lookup?

Start by reading the full statement, solve the visible examples, then compare the pattern against related Persona online assessment problems in FastPrep.

Is In-Memory Database with TTL and Historical Lookup associated with Persona?

FastPrep tags this as a Persona online assessment coding practice problem when the problem catalog has that company and stage context.

In-Memory Database with TTL and Historical Lookup

Persona● HardINTERNOA

Learn

Problem statement

Implement an in-memory database that stores integer values under string record keys and string field names. Support conditional changes, sorted scans, per-field expiration, and historical reads. All four levels are combined in this exercise; earlier operations remain available at later levels.

Complete runHistoricalDatabase. Process operations in order and return one string array for each operation. Each row begins with an operation name and a strictly increasing timestamp. Numeric arguments are represented as decimal strings.

Level 1: Values and conditional changes

["SET", timestamp, key, field, value]: create or replace a field with a non-expiring value. Return [].
["GET", timestamp, key, field]: return [value] if the field is visible, otherwise [].
["COMPARE_AND_SET", timestamp, key, field, expectedValue, newValue]: if the visible value equals expectedValue, replace it with the non-expiring newValue and return ["true"]. Otherwise return ["false"] without changing anything.
["COMPARE_AND_DELETE", timestamp, key, field, expectedValue]: delete the field only if its visible value equals expectedValue. Return ["true"] on success and ["false"] otherwise.

Level 2: Ordered scans

["SCAN", timestamp, key]: return all visible fields of the record as strings field(value), ordered lexicographically by field name.
["SCAN_BY_PREFIX", timestamp, key, prefix]: return only visible fields whose names begin with the case-sensitive prefix, in the same format and order.
A missing record, no matching fields, or a record with only expired or deleted fields produces [].

Level 3: Time to live

["SET_WITH_TTL", timestamp, key, field, value, ttl]: replace the field with value, visible on the half-open interval [timestamp, timestamp + ttl). Return [].
["COMPARE_AND_SET_WITH_TTL", timestamp, key, field, expectedValue, newValue, ttl]: perform the same visible-value comparison as COMPARE_AND_SET. On success the new value expires at timestamp + ttl; return ["true"]. A failed comparison changes neither value nor expiration and returns ["false"].
Every current read, scan, comparison, and deletion treats an expired field as absent. Successful non-TTL writes remove any previous TTL.

Level 4: Historical lookup

["GET_WHEN", timestamp, key, field, atTimestamp]: return the value visible at atTimestamp as [value], or [] if no value was visible then.
If atTimestamp is 0, perform a current GET at timestamp.
Overwrites and successful deletions take effect at their timestamp but preserve earlier history. A historical read at an exact mutation timestamp observes that mutation.
Expiration never restores an older overwritten value. A read first considers the most recent successful mutation at or before the requested time; if that version is deleted or expired, return [].
Failed conditional operations create no version and have no effect on later historical reads.

Function

runHistoricalDatabase(operations: String[][]) → String[][]

Examples

Example 1

operations = [["SET_WITH_TTL","1","A","B","3","10"],["COMPARE_AND_SET_WITH_TTL","4","A","B","3","7","9"],["GET","10","A","B"],["GET_WHEN","13","A","B","3"],["GET_WHEN","15","A","B","13"]]return = [[],["true"],["7"],["3"],[]]

The first value is visible from time 1 until the overwrite at 4. The replacement is visible on [4,13). The historical read at 3 returns 3; the read at its exact expiration time 13 returns no value.

Example 2

operations = [["SET","1","user","age","18"],["SET","2","user","address","1"],["SET","3","user","height","180"],["SCAN","4","user"],["SCAN_BY_PREFIX","5","user","a"],["COMPARE_AND_DELETE","6","user","age","18"],["GET","7","user","age"],["GET_WHEN","8","user","age","5"]]return = [[],[],[],["address(1)","age(18)","height(180)"],["address(1)","age(18)"],["true"],[],["18"]]

Scans sort by field name; the prefix scan selects address and age. The successful deletion hides the current age but does not erase its earlier value.

Constraints

1 <= operations.length <= 2000.
Each row has exactly the arguments shown for its operation.
Operation timestamps are integers in [1, 10^9] and strictly increase.
0 <= atTimestamp <= timestamp.
1 <= ttl <= 10^9.
Values and expected values are canonical signed 32-bit decimal integers.
Keys, fields, and prefixes contain 1 through 50 case-sensitive English letters or digits.
The database is empty before the first operation.