Operator List

The list below shows all currently known operators, with their icons as they appear in various tools. The first column shows the icon for each operator as used in the current version of Microsoft’s querying tools: SQL Server Management Studio (SSMS) and in Azure Data Studio (ADS). The second column shows the legacy artwork, as used in SSMS versions 17.3 and older (before December 2017). The third column shows the icons that are used in SolarWinds’ SQL Serntry Plan Explorer, the most popular third-party product for working with execution plans. (Note that until February 2019, Plan Explorer used the same icons as Legacy SSMS). And finally, in the fourth column, you’ll find the icons as used on Brent Ozar’s Paste The Plan, a website designed for easy sharing of execution plans on media that don’t allow attachments but do allow links.

The operator names are hyperlinks to their detailed description. If a name is not a hyperlink, there is no detailed description page for that operator yet.

Name	Legacy SSMS	Paste The Plan	Short description
Adaptive Join			Joins two inputs using either hashing or nested loops. The choice is made at run-time, depending on the number of rows in the first input.
Assert			Verifies a condition; raises an error and aborts processing if the condition is met.
Batch Hash Table Build			Maps input values for a join to a hash table, that can then be used to reduce the number of rows to be processed for the other input of that join. Only on SQL Server 2012.
Bitmap			Maps input values for a join to a bitmap, that can then be used to reduce the number of rows to be processed for the other input of that join.
Catchall			This icon is used when no suitable specific icon for the operation exists. It is mostly seen for cursor operations. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Clustered Index Delete			Deletes rows from a clustered index, and optionally from one or more nonclustered indexes at the same time.
Clustered Index Insert			Inserts rows into a clustered index, and optionally into one or more nonclustered indexes at the same time.
Clustered Index Merge			Applies a stream of insert, update, and delete instructions to a clustered index, and optionally applies the corresponding changes to one or more nonclustered indexes at the same time.
Clustered Index Scan			Reads all data from a clustered index, either in allocation order or in logical (index) order.
Clustered Index Seek			Finds a specific row in a clustered index, based on key value; and optionally continues to scan from there in logical (index) order.
Clustered Index Update			Updates rows in a clustered index, and optionally in one or more nonclustered indexes at the same time.
Collapse			Optimizes update processing by replacing a delete and insert for the same key value with a single update.
Columnstore Index Delete		(shown as Index Delete or Clustered Index Delete)	Deletes rows from a columnstore index.
Columnstore Index Insert		(shown as Index Insert or Clustered Index Insert)	Inserts rows into a columnstore index.
Columnstore Index Merge		(shown as Clustered Index Merge)	Applies a stream of insert, update, and delete instructions to a columnstore index.
Columnstore Index Scan		(shown as Index Scan or Clustered Index Scan)	Reads all data from a columnstore index.
Columnstore Index Update		(shown as Index Update or Clustered Index Update)	Updates rows in a columnstore index.
Compute Scalar			Evaluates one or more expressions and adds their results to the data stream.
Concatenation			Processes two or more inputs, in order, returning all rows returned from each of those inputs.
COND			Represents the condition evaluation of IF and WHILE statements. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Constant Scan			Generates one or more rows with zero or more columns, with constant values as defined in the operator’s properties.
Deleted Scan		(no icon)	Reads all data from the deleted pseudo-table. Used in triggers only.
Dynamic			Represents the declaration of a dynamic cursor. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Fetch Query			Represents the query to be used for FETCH statements. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Filter			Applies a predicate to its input rows and returns only rows for which the predicate evaluates to True.
Foreign Key References Check	(not supported)	(no icon)	Performs referential integrity checking for modified data in cases where more than 253 foreign key references exist on a single primary key or unique constraint. Only on SQL Server 2016 and newer.
Hash Match			Builds a hash table from its first input, then uses that hash table to either join to its second input, or produce aggregated values.
Index Delete			Deletes rows from a nonclustered index.
Index Insert			Inserts rows into a nonclustered index.
Index Scan			Reads all data from a nonclustered index, either in allocation order or in logical (index) order.
Index Seek			Finds a specific row in a nonclustered index, based on key value; and optionally continues to scan from there in logical (index) order.
Index Spool			Stores its input rows in an internal, indexed worktable; this indexed worktable can then be used to re-process specific subsets of the data.
Index Update			Updates rows in a nonclustered index; may also be used to insert and/or delete rows from that same index.
Inserted Scan			Reads all data from the inserted pseudo-table. Used in triggers only.
Key Lookup			Reads a single row from a clustered index, based on a key value that was retrieved from a nonclustered index on the same table.
Keyset			Represents the declaration of a keyset cursor. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Language Element			This icon is used for operators for which no suitable specific icon is available. It is mostly seen as the top-left operator of an execution plan for operations other than `SELECT`, serving as the container for properties of the plan as a whole. It can also be used, in an execution plan only, for statements with no execution plan.
Merge Interval			Combines multiple sets of potentially overlapping intervals into lesser but non-overlapping intervals.
Merge Join			Joins two inputs that are ordered by the join key(s), exploiting the known sort order for optimal processing efficiency.
Nested Loops			Joins two inputs by repeatedly executing the second input for each row in the first input.
Online Index Insert			For an online index rebuild, this operator represents the insertion of data in an index while keeping the index online.
Parallelism (Distribute Streams)			The parallelism operator, also known as exchange operator, manages the distribution of rows between threads in parallel plans. The icon changes depending on the logical operation, and the functionality changes sufficiently that each logical operation is described separately. The Distribute Streams logical operation receives rows from a serial region of the plan, and distributes them to multiple threads in a parallel region of the plan.
Parallelism (Gather Streams)			The parallelism operator, also known as exchange operator, manages the distribution of rows between threads in parallel plans. The icon changes depending on the logical operation, and the functionality changes sufficiently that each logical operation is described separately. The Gather Streams logical operation gathers rows from the multiple threads of a parallel region of the plan, and sends them to a serial region of the plan.
Parallelism (Repartition Streams)			The parallelism operator, also known as exchange operator, manages the distribution of rows between threads in parallel plans. The icon changes depending on the logical operation, and the functionality changes sufficiently that each logical operation is described separately. The Repartition Streams logical operation gathers rows from the multiple threads of a parallel region of the plan, and distributes them to multiple threads in another parallel region of the plan, where a different distribution is needed.
Parameter Table Scan			Used to insert the results of remotely executed code into a local table.
Population Query			Represents the query to be used for the OPEN cursor statement. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Put	(not supported)	(no icon)	Used to insert data in external (Polybase) tables. Only on SQL Server 2016 and newer.
Remote Delete			Deletes rows from a remote object.
Remote Insert			Inserts rows into a remote object.
Remote Query			Submits a query to a remote destination for execution there. The operator will then return all rows produced by the remote destination.
Remote Scan			Reads all data from a remote object.
Remote Update			Updates rows in a remote object.
Result			Also known as “Select”, this icon is used as the top-left operator in any plan for a `SELECT` statement. It can be considered the representation of the client, and serves as the container for properties of the plan as a whole.
RID Lookup			Reads a single row from a heap table, based on an RID value that was retrieved from a nonclustered index on the same table.
Row Count Spool			Counts the number of rows in its input and then returns the same number of empty rows.
Segment			Marks the start of each new segment in the data stream by inspecting specific columns for a change in their data.
Sequence			Processes two or more inputs, in order, returning only rows returned from the final input.
Sequence Project			Computes values that are determined by a row’s relative position within the entire data stream.
Snapshot			Represents the declaration of a static cursor. Only in execution plan only; not in execution plan plus run-time statistics or live execution plan.
Sort			Reads all input rows, sorts them, and then returns them in the specified order.
Split			Optimizes update processing by replacing update operations in the corresponding delete and insert operations.
Stream Aggregate			Computes aggregation results by reading a sorted input stream and returning a single row for each set of rows with the same key value.
Switch			Executes only one of its inputs; which input is executed is determined at runtime.
Table Delete			Deletes rows from a heap table, and optionally from one or more nonclustered indexes at the same time.
Table Insert			Inserts rows into a heap table, and optionally into one or more nonclustered indexes at the same time.
Table Merge		(no icon)	Applies a stream of insert, update, and delete instructions to a heap table, and optionally applies the corresponding changes to one or more nonclustered indexes at the same time.
Table Scan			Reads all data from a heap table, in allocation order.
Table Spool			Stores its input rows in an internal worktable; this worktable can then be used to re-process the same data.
Table Update			Updates rows in a heap table, and optionally in one or more nonclustered indexes at the same time.
Table Valued Function			Executes a table-valued function and stores the results in a temporary table.
Top			Returns only the first number or percent of its input rows, optionally skipping the first few and optionally adding extra rows based on ties in specified columns.
UDX			A catch-all operator icon and name for what is in fact a set of different “extended” operators that are involved in various XML (XQuery and XPath) and JSON operations.
Window Aggregate			Evaluates one or more aggregate expressions with a window definition and adds their results to the data stream.
Window Spool			Stores its input rows in an internal worktable; for each row then returns all rows associated with its window specification.