Build a Pivot Table Aggregator

Implement a function that builds a pivot table from a list of records. Each record contains a row_key, a column_key, and a numeric value.

Requirements:

1. In the basic version, aggregate records by (row_key, column_key) and compute the sum for each group.
2. Return the result in a structure that allows efficient lookup by row key and column key.
3. As a follow-up, explain how you would extend the implementation to support other aggregation functions such as average and minimum.

Example input:

[
  ("A", "X", 10),
  ("A", "X", 20),
  ("A", "Y", 5),
  ("B", "X", 7)
]

Expected result for sum:

• A / X -> 30
• A / Y -> 5
• B / X -> 7

In your explanation, describe what intermediate state must be maintained so that average and minimum can also be computed correctly.

Part 1: Build a Sum-Based Pivot Table Aggregator

You are given a list of records. Each record is a tuple of the form (row_key, column_key, value). Build a pivot table that groups records by the pair (row_key, column_key) and stores the sum of all values in that group.

Return the result as a nested dictionary so that lookups are efficient: result[row_key][column_key] should give the aggregated sum for that cell.

Do not create entries for missing row/column combinations. If the input is empty, return an empty dictionary.

Constraints

• 0 <= len(records) <= 200000
• row_key and column_key are strings
• value is an integer or float
• Only row/column pairs that appear in the input should appear in the output

Examples

Example 1

Input: ([('A', 'X', 10), ('A', 'X', 20), ('A', 'Y', 5), ('B', 'X', 7)],)
Output: {'A': {'X': 30, 'Y': 5}, 'B': {'X': 7}}

Explanation: The pair ('A', 'X') appears twice, so its values are added: 10 + 20 = 30.

Example 2

Input: ([],)
Output: {}

Explanation: Edge case: no records means no pivot table entries.

Hints

1. A hash map (dictionary) is a natural fit for fast grouping and lookup.
2. Think of the result as a dictionary of dictionaries: first by row, then by column.

Part 2: Extend the Pivot Table Aggregator to Support Sum, Average, and Minimum

You are given a list of records. Each record is a tuple of the form (row_key, column_key, value). Implement a pivot table aggregator that groups records by (row_key, column_key) and supports three aggregation modes:

• "sum": sum all values in each cell
• "average": compute the arithmetic mean of values in each cell
• "minimum": keep the smallest value in each cell

Return the result as a nested dictionary so that result[row_key][column_key] gives the aggregated value.

Important design idea: different aggregation functions require different intermediate state. For example, average cannot be computed correctly from just the latest value; it needs both a running total and a count. Minimum needs the smallest value seen so far for each cell.

If the input is empty, return an empty dictionary. You may assume agg_func is always one of "sum", "average", or "minimum".

Constraints

• 0 <= len(records) <= 200000
• row_key and column_key are strings
• value is an integer or float
• agg_func is one of: "sum", "average", "minimum"
• Only row/column pairs that appear in the input should appear in the output

Examples

Example 1

Input: ([('A', 'X', 10), ('A', 'X', 20), ('A', 'Y', 5), ('B', 'X', 7)], 'average')
Output: {'A': {'X': 15.0, 'Y': 5.0}, 'B': {'X': 7.0}}

Explanation: For ('A', 'X'), average = (10 + 20) / 2 = 15.0.

Example 2

Input: ([('A', 'X', 10), ('A', 'X', 20), ('A', 'Y', 5), ('B', 'X', 7)], 'minimum')
Output: {'A': {'X': 10, 'Y': 5}, 'B': {'X': 7}}

Explanation: For each cell, keep the smallest value seen.

Hints

1. For average, each cell needs more than one number while processing. What two pieces of information are enough?
2. You can maintain one nested dictionary for state while scanning the records, then convert that state into the final pivot table.