DP-750 Certification Practice Question #43

Question

Match each scenario on the left to the most appropriate table/layout approach on the right. Each tile is used exactly once.

```mermaid
flowchart LR
  subgraph Scenarios
    A["1. New gold table; want lowest cost,
auto-optimization (predictive optimization,
auto-compaction), latest UC features"]
    B["2. External Spark/Trino writers manage the
data files in cloud storage outside Databricks;
Databricks only governs access"]
    C["3. Large new fact table queried by changing,
high-cardinality filters; want self-tuning
layout with no re-OPTIMIZE column re-spec"]
    D["4. Migrating non-Delta Parquet data already
in storage during a Hive-to-UC upgrade
without moving the files"]
  end
  subgraph Tiles
    T1["Unity Catalog managed Delta table"]
    T2["External table (data lifecycle
managed outside Databricks)"]
    T3["Liquid clustering (CLUSTER BY)"]
    T4["External Parquet table
(register in place, no move)"]
  end
```

Accepted Answer

Microsoft recommends Unity Catalog managed Delta tables for all new tables to get auto-compaction, predictive optimization, and the latest features (Scenario 1). External tables are the right choice when external engines own the data lifecycle/file layout and Databricks only governs access (Scenario 2), and they are also required to register non-Delta formats such as in-place Parquet during a Hive-to-UC migration without moving files (Scenario 4). Liquid clustering is the self-tuning layout for large tables filtered by changing, high-cardinality columns, persisting its keys so you don't re-specify columns on each `OPTIMIZE` (Scenario 3).

More DP-750 practice questions