DP-750 Certification Practice Question #57

Question

You must build a silver aggregate that answers: **"For fulfilled orders only, how many orders exist per order priority, sorted from most to fewest orders?"** The source is a DataFrame `df_order` with columns `o_orderkey`, `o_orderstatus`, and `o_orderpriority`. A fulfilled order has `o_orderstatus == "F"`.

You want a single chained PySpark expression that filters to fulfilled orders, groups by priority, counts the orders, names the count column `n_orders`, and sorts descending by `n_orders`.

```python
from pyspark.sql.functions import col, count

df_chained = (
    df_order
    # filter -> group -> aggregate -> sort
)
```

Which expression produces the required result?

Accepted Answer

The correct lazy-evaluation chain is filter → groupBy → agg → sort. Filtering to `o_orderstatus == "F"` must happen before the `groupBy`, because after grouping by `o_orderpriority` the `o_orderstatus` column is no longer available to filter on. The aggregation uses `count(...).alias("n_orders")` so the result column can be sorted, and `sort(col("n_orders").desc())` orders from most to fewest. Option C produces ascending order with an unaliased `count` column, and options A and D place stages in an invalid order.

More DP-750 practice questions