The Python Round

def flatten(obj, prefix="", sep="."):
    out = {}
    if isinstance(obj, dict):
        for k, v in obj.items():
            key = f"{prefix}{sep}{k}" if prefix else k
            out.update(flatten(v, key, sep))
    elif isinstance(obj, list):
        out[prefix] = ",".join(str(x) for x in obj)
    else:
        out[prefix] = obj
    return out

# Example
nested = {"user": {"id": 1, "addr": {"city": "NYC", "zip": "10001"}}}
print(flatten(nested))
# {"user.id": 1, "user.addr.city": "NYC", "user.addr.zip": "10001"}

def dedup_latest(records):
    latest = {}
    for r in records:
        key = (r["user_id"], r["event_type"])
        if key not in latest or r["ts"] > latest[key]["ts"]:
            latest[key] = r
    return list(latest.values())

# Edge cases to mention:
# - Empty input returns []
# - Equal timestamps: keeps whichever comes last in input order

from datetime import timedelta

def sessionize(events, gap_minutes=30):
    events = sorted(events, key=lambda e: (e["user_id"], e["ts"]))
    threshold = timedelta(minutes=gap_minutes)
    out = []
    last_user = None
    last_ts = None
    session_id = 0
    for e in events:
        if e["user_id"] != last_user or (e["ts"] - last_ts) > threshold:
            session_id += 1
        out.append({**e, "session_id": session_id})
        last_user, last_ts = e["user_id"], e["ts"]
    return out

import csv
from typing import Iterator

def stream_clean_csv(path: str, chunk_size: int = 10_000) -> Iterator[list[dict]]:
    with open(path, newline="") as f:
        reader = csv.DictReader(f)
        chunk = []
        for row in reader:
            if not row.get("email"):
                continue
            row["email"] = row["email"].lower().strip()
            chunk.append(row)
            if len(chunk) >= chunk_size:
                yield chunk
                chunk = []
        if chunk:
            yield chunk

for batch in stream_clean_csv("users.csv"):
    bulk_insert(batch)  # production: writes to Postgres or S3

def inner_join(left, right, key):
    right_index = {}
    for r in right:
        right_index.setdefault(r[key], []).append(r)
    out = []
    for l in left:
        for r in right_index.get(l[key], []):
            merged = {**l, **r}
            out.append(merged)
    return out

# Note: right value overwrites left on key collision.
# Mention this and ask the interviewer if it is acceptable.