Token Consumption Moving Averages | Dataford Interview Questions

Your question is Token Consumption Moving Averages. Start with the requirements and the two tables on the right.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem

Anthropic needs a daily view of token consumption for enterprise accounts using the Claude API. Write a PostgreSQL query that aggregates usage events and calculates both short-term usage trends and cumulative consumption.

Requirements

Include only accounts where account_type = 'enterprise' and status = 'active'.
Aggregate multiple token usage events for the same account and date into daily_tokens. Treat NULL token counts as zero.
Calculate a three-row moving average, including the current day and up to the two preceding observed usage days for that account.
Calculate a cumulative token total for each account in chronological order, and sort the final output by account name and usage date.

Schema

enterprise_accounts

Column	Type	Description
account_idPK	INTEGER	Unique enterprise account identifier
account_name	VARCHAR(100)	Enterprise customer name
account_type	VARCHAR(30)	Customer segment
status	VARCHAR(20)	Account lifecycle status

token_usage

Column	Type	Description
usage_idPK	BIGINT	Unique token usage event identifier
account_id	INTEGER	References enterprise_accounts.account_id
usage_date	DATE	UTC calendar date of token usage
token_count	BIGINT	Tokens consumed in the event