How should I practice Record Linkage Part 3 - Full Connected Component?

Start by reading the full statement, solve the visible examples, then compare the pattern against related Stripe phone screen problems in FastPrep.

Is Record Linkage Part 3 - Full Connected Component associated with Stripe?

FastPrep tags this as a Stripe phone screen coding practice problem when the problem catalog has that company and stage context.

Problem · Union Find

Record Linkage Part 3 - Full Connected Component

Learn this problem

● Medium

StripeFULLTIMEPHONE SCREEN

See Stripe hiring insights

Problem statement

This continues the record-linkage problem. You are given user records and a weighted similarity rule, and must return the full set of records transitively connected to a target.

Inputs: rows (a String[] of "id,name,email,company"), weights (a String[] of "field,weight" over {name, email, company}, summing to 1), threshold, and targetId.

Two records are linked when their weighted field-equality similarity is >= threshold (each matching field contributes its weight). Linkage is transitive: if A links to B and B links to C, then A, B, and C are all in the same group even if A and C are not directly linked.

Return the ids of every record in the same connected component as targetId (i.e. reachable through any chain of links), excluding the target itself, as an int[] sorted in ascending order.

Function

findLinkedComponent(rows: String[], weights: String[], threshold: float, targetId: int) → int[]

Examples

Example 1

rows = ["1,Alice,alice@gmail.com,Stripe", "2,Alicia,alice@gmail.com,Stripe", "3,Alice,alice@yahoo.com,Google", "4,Bob,bob@gmail.com,Stripe"]weights = ["name,0.2", "email,0.5", "company,0.3"]threshold = 0.5targetId = 1return = [2]

Record 1 links directly to record 2 (similarity 0.8). Record 2 does not link to record 3 (0) or record 4 (0.3 < 0.5), and record 1 links to neither 3 nor 4. So the component of record 1 is {1, 2}; excluding the target, the answer is [2].

Example 2

rows = ["1,Alice,alice@gmail.com,Stripe", "2,Alicia,alice@gmail.com,Stripe", "3,Bob,bob@yahoo.com,Google", "5,Alicia,carol@outlook.com,Stripe"]weights = ["name,0.2", "email,0.5", "company,0.3"]threshold = 0.5targetId = 1return = [2, 5]

Record 1 links to record 2 (email + company match = 0.8). Record 2 links to record 5 (name 'Alicia' + company 'Stripe' match = 0.2 + 0.3 = 0.5 >= threshold), but record 5 does NOT link directly to record 1 (only company matches = 0.3 < 0.5). Transitivity pulls record 5 into record 1's component via record 2. Record 3 links to nobody. Component of 1 = {1, 2, 5}; excluding the target, the answer is [2, 5].

Constraints

Each row is "id,name,email,company" with an integer id.
Each weight entry is "field,weight" over {name, email, company}; weights sum to 1.
Similarity = sum of weights of matching fields; records are linked when similarity >= threshold.
Linkage is transitive (connected components over the similarity graph).
Return the ascending-sorted int[] of ids in the target's component, excluding the target.