Skip to main content
Aggregate functions let you reference grand totals and row totals in your table calculations. Unlike row functions and pivot functions which compile to SQL window functions, aggregate functions generate separate CTEs that compute totals from the underlying data. This means total() correctly handles all metric types — including count_distinct and average — by re-running the metric’s native aggregation rather than naively summing grouped values.
Aggregate functions only accept metric field references — dimensions and other table calculations are not valid arguments. Passing a dimension (e.g. total(${orders.customer_id})) returns the error “Tried to reference metric with unknown field id”.

​
total

Returns the grand total of a metric across all rows, computed from the raw data using the metric’s native aggregation.
Example Calculate each row’s share of total revenue:
With percent formatting applied, this gives you each row as a percentage of the overall total. Edit Table Calculation dialog with a percent_of_total formula using total()
Lightdash generates a column_totals CTE that re-aggregates the metric from the raw data with no GROUP BY, producing a single grand-total value. This value is then cross-joined into the main query.For a sum metric, the CTE computes SUM(...). For count_distinct, it computes COUNT(DISTINCT ...). For average, it computes AVG(...). Each metric type uses its own native aggregation, so the total is always mathematically correct.

​
row_total

Returns the sum of a metric’s values across all pivot columns for the current row.
row_total is only available when your query includes a pivoted dimension. If no pivot is configured, row_total(${table.metric}) falls back to the metric’s value directly.
Example In a pivot table with revenue broken out by region, calculate each region’s share of the row total:
Lightdash generates a row_totals CTE that reads from the already-grouped results and computes SUM(metric) grouped by the non-pivot dimensions. This gives one total per row, which is then joined back into the main query.Unlike total(), row_total() always uses SUM regardless of the metric type, since it’s summing pre-aggregated values across the pivot columns within each row.