Apache DataFusion SQL Query Engine
doc: add comments for `LocalLimitExec` and `GlobalLimitExec` (#25993)
Yongting You22a0fdb64614
diff --git a/datafusion/physical-plan/src/limit.rs b/datafusion/physical-plan/src/limit.rsindex b9020a512b..ef504a40ea 100644--- a/datafusion/physical-plan/src/limit.rs+++ b/datafusion/physical-plan/src/limit.rs@@ -45,7 +45,20 @@ use datafusion_physical_expr::{LexOrdering, PhysicalExpr};use futures::stream::{Stream, StreamExt};use log::trace;-/// Limit execution plan+/// Applies a limit to the entire result.+///+/// Requires a single input partition; otherwise, plan validation or execution fails.+/// For input with multiple partitions, merge them before applying the limit:+/// (For SQL interface, coalescing will be enforced automatically in physical+/// optimizer through [`ExecutionPlan::required_input_distribution`])+///+/// ```text+/// GlobalLimitExec: skip=0, fetch=10+/// CoalescePartitionsExec <--- combine to 1 partition output+/// DataSourceExec: partitions=4 <--- 4 parallel scan partitions+/// ```+///+/// To enforce a per-partition limit, use [`LocalLimitExec`].#[derive(Debug, Clone)]pub struct GlobalLimitExec {/// Input execution plan@@ -352,7 +365,17 @@ impl GlobalLimitExec {}}-/// LocalLimitExec applies a limit to a single partition+/// Applies a limit independently to each input partition.+///+/// Preserves input partitioning. With four partitions, a limit of 10 allows+/// up to 40 rows in total:+///+/// ```text+/// LocalLimitExec: fetch=10+/// DataSourceExec: partitions=4+/// ```+///+/// To enforce a limit on the entire result, use [`GlobalLimitExec`].#[derive(Debug, Clone)]pub struct LocalLimitExec {/// Input execution plan