Given sales(sale_id, user_id, amount), return the top 2 rows for each user_id by amount descending (ties broken by sale_id ascending). Project user_id, sale_id, amount. Sort by user_id asc, amount desc, sale_id asc.
| sale_id | user_id | amount |
|---|---|---|
| 1 | 2 | 353.58 |
| 2 | 4 | 209.65 |
| 3 | 2 | 69.15 |
| 4 | 1 | 194.07 |
| 5 | 2 | 200.94 |
| 6 | 3 | 479.28 |
These DataFrames are passed to solve(spark, dfs). Access one with dfs['name'].
sales12 rows| sale_id (int) | user_id (int) | amount (double) |
|---|---|---|
| 1 | 2 | 353.58 |
| 2 | 4 | 209.65 |
| 3 | 2 | 69.15 |
| 4 | 1 | 194.07 |
| 5 | 2 | 200.94 |
| 6 | 3 | 479.28 |
| 7 | 3 | 92.03 |
| 8 | 3 | 446.03 |
| 9 | 4 | 161.2 |
| 10 | 3 | 478.33 |
| 11 | 2 | 429.8 |
| 12 | 2 | 201.56 |
Column names and order must match. Sort order is specified in the question.