Think about which function accumulates values over a range.
RUNNING_SUM accumulates the sum of sales in order, producing a running total. TOTAL returns the total for the partition, WINDOW_SUM sums over a window but needs specific offsets, and SUM just sums the current row.
Think about which chart type shows trends over time smoothly.
A line chart is ideal for showing cumulative trends over time because it connects points smoothly and shows growth clearly. Bar charts and pie charts do not show accumulation well.
Partitioning controls where the running total resets.
Without partitioning by region, the running total sums sales across all regions as one group, accumulating over time without resetting.
Ordering affects how running totals accumulate.
If the running total is ordered by sales amount rather than date, the calculation resets unexpectedly because the sequence is not chronological.
Context filters affect calculation order and what data is included.
To ignore product category filters but respect region filters, set region filters as context filters and exclude product category filters from context. This way, running total respects region but ignores product category filtering.