Blocking operator in an execution plan is an operator that pauses the flow of data during query execution. This behaviour can impact performance as it delays subsequent operations in the query pipeline, which must wait for the blocking operator to complete its task before they can proceed. Blocking operators are often associated with operations that require data to be fully sorted, aggregated, or reorganized before moving forward.
Characteristics of Blocking Operators
Ø Stops Row Flow: These operators require all data to be gathered or processed before passing it downstream.
Ø Common in Aggregations and Sorting: Typically found in operators that need to reorder or aggregate large datasets, such as Sort, Hash Match, and Sort Aggregate.
Ø Potentially High Memory Usage: Since they need to process a lot of data at once, blocking operators often use more memory and can lead to temporary disk storage (spooling) if memory is insufficient.
Common Blocking Operators in SQL Server
Here are some typical blocking
operators:
Ø Sort Operator: Requires data to be fully sorted before passing it on, blocking subsequent operations.
Ø Hash Match (Aggregate) Operator: Used in hash joins and aggregations, blocks the flow until all rows are processed.
Ø Parallelism (Distribute Streams): In a parallel execution plan, distributes data to parallel streams, potentially causing delays.
Ø Spool Operators: Operators like Table Spool and Index Spool can also act as blocking operators, especially if they must store intermediate results before continuing.
