Dataset Validation (CSV with banned + stop words)

Tech-screen string-processing round. Input is a multi-line CSV string. Part 1 validates non-empty fields, Part 2 filters banned words, Part 3 tokenizes specific columns and counts stop-word occurrences. Requirements Input is a single multi-line CSV string. Read it row by row (one report used Java's BufferedReader). Part 1: validate that each row has no empty string fields; report or drop invalid rows per the spec. Part 2: given a list of banned words, strip / mask them out of the affected text columns. Part 3: given a set of stop words, tokenize two designated columns and count total stop-word occurrences across those columns. Notes Prompt is reported as long and tricky to parse; reading-comprehension speed is the bottleneck. Interviewer in the source report was helpful and pulled out the requirements when asked. This is essentially a string-handling stress test — no real algorithms involved. Preparation Pre-write CSV-by-hand parsers in your language of choice (no pandas). Practice tokenizing with re.findall(r"\\w+", ...) and case-folding. Drill on writing 3-part prompts with shared input under 45 minutes — the time risk is reading, not coding.