Query Plan
Understand what cypher is doing
Cypher breaks down the work of executing a query into small pieces called operators. Each operator is responsible for a small part of the overall query. The operators are connected together in a pattern called a Query Plan.
When you use the EXPLAIN or PROFILE commands, Neo4j Browser displays a diagram of the Query Plan.
Click the right arrow button for the later pages of this guide, which explain how to read the Query Plan diagram.
Operators
Each Operator is displayed as a rectangle with its name in the top-left corner. See the operators manual page for a description of what each operator does.
Pipes
Operators are connected by pipes; each pipe represents the output of one operator and the input of the next.
Rows
In Query Plan diagrams, the width of a pipe indicates the number of rows that pass between operators. This means that by looking at the overall diagram, you can quickly see the shape of the query, in terms of which parts process lots of rows, and which parts process few.
Estimated rows
When you use the EXPLAIN keyword, the query is not actually executed; so it's not possible to show actual number of rows for each pipe. In this case, the Query Plan diagram shows estimated rows instead. These numbers are predicted based on Neo4j's built-in statistics. The Cypher cost-based planner uses estimated rows to determine the optimal query plan.
Database hits
Each operator will ask the Neo4j storage engine to do work such as retrieving or updating data. A database hit is an abstract unit of this storage engine work.
When you use the PROFILEcommand, the footer of the result frame displays the total number of database hits incurred while running the query. By comparing this total for different query plans, you can tell which one is better in terms of work for the storage engine.
In addition of the total number of database hits, the Query Plan diagram shows the database hits for each individual operator. For operators with a large number of database hits, there is a red box underneath the operator, with its height proportional to the number. This means that you can glance very quickly at the whole Query Plan diagram and spot the operators that are responsible for significant storage engine work.
Click to expand
Some operators can reveal more information about what they are doing. If you see a triangular expand icon next to the operator name, you can click to expand the operator and reveal some more details. When you're done, you can click the header again to collapse the operator and hide the details.
If you want to quickly expand all the operators, there are expand-all and collapse-all buttons below the diagram.
Identifiers
The first section of details lists the identifiers that the operator works with.
if you've named identifers in your query, e.g. MATCH (n) then you should find the identifier n bound in one of the operators.
In addition to identifiers that you named in your query you may see
some internal identifiers that Cypher has introduced to keep track
of unnamed entities.
Expressions
After the identifiers, you'll see a section for an expression.
This is commonly a boolean expression such as hasProp(born) or it can be a graph pattern to be expanded such as ()-[r]-()
| Reference: | Execution Plans manual page |
|---|---|
| Related: | :help EXPLAIN :help PROFILE |