แก้ไข

แชร์ผ่าน


Azure Digital Twins query language reference: MATCH clause

This document contains reference information on the MATCH clause for the Azure Digital Twins query language.

The MATCH clause is used in the Azure Digital Twins query language as part of the FROM clause. MATCH allows you to specify which pattern should be followed while traversing relationships in the Azure Digital Twins graph (this is also known as a "variable hop" query pattern).

This clause is optional while querying.

Core syntax: MATCH

MATCH supports any query that finds a path between twins within a range of hops, based on certain relationship conditions.

The relationship condition can include one or more of the following details:

A query with a MATCH clause must also use the WHERE clause to specify the $dtId for at least one of the twins it references.

Note

MATCH is a superset of all JOIN queries that can be performed in the query store.

Syntax

Here's the basic MATCH syntax.

It contains these placeholders:

--SELECT ... FROM ...  
MATCH (twin_or_twin_collection_1)-[relationship_condition]-(twin_or_twin_collection_2)
WHERE twin_or_twin_collection.$dtId = 'twin_ID' 
-- AND ...

You can leave one of the twin collections blank in order to allow any twin to work in that spot.

You can also change the number of relationship conditions, to have multiple chained relationship conditions or no relationship condition at all:

--Chained relationship conditions
-- SELECT ... FROM ... 
MATCH (twin_or_twin_collection_1)-[relationship_condition]-(twin_or_twin_collection_2)-[relationship_condition]-(twin_or_twin_collection_3)...
WHERE twin_or_twin_collection.$dtId = 'twin_ID'
-- No relationship condition
-- SELECT ... FROM ... 
MATCH (twin_or_twin_collection_1)
WHERE twin_or_twin_collection.$dtId = 'twin_ID'

For more detail about each type of relationship condition and how to combine them, see the other sections of this document.

Example

Here is an example query using MATCH.

The query specifies a relationship direction, and searches for Building and Sensor twins where...

  • the Sensor is targeted by any relationship from a Building twin with a $dtId of Building21, and
  • the Sensor has a temperature above 50. The Building and Sensor are both included in the query result.
SELECT Building, Sensor FROM DIGITALTWINS 
MATCH (Building)-[]->(Sensor) 
WHERE Building.$dtId = 'Building21' AND Sensor.temperature > 50

Specify relationship direction

Use the relationship condition in the MATCH clause to specify a relationship direction between the twins. Possible directions include left-to-right, right-to-left, or non-directional. Cyclic relationships are automatically detected, so that a relationship is traversed only once.

Note

It's possible to represent bi-directional relationships by using chaining.

Syntax

Note

The examples in this section focus on relationship direction. They don't specify relationship names, they default to a single hop, and they don't assign query variables to the relationships. For instructions on how to do more with these other conditions, see Specify relationship name, Specify number of hops, and Assign query variable to relationship. For information about how to use several of these together in the same query, see Combine MATCH operations.

Directional relationship descriptions use a visual depiction of an arrow to indicate the direction of the relationship. The arrow includes a space set aside by square brackets ([]) for an optional relationship name.

This section shows the syntax for different directions of relationships. The placeholder values that should be replaced with your values are source_twin_or_twin_collection and target_twin_or_twin_collection.

For a left-to-right relationship, use the following syntax.

-- SELECT ... FROM ...
MATCH (source_twin_or_twin_collection)-[]->(target_twin_or_twin_collection)
-- WHERE ...

For a right-to-left relationship, use the following syntax.

-- SELECT ... FROM ...
MATCH (target_twin_or_twin_collection)<-[]-(source_twin_or_twin_collection)
-- WHERE ...

For a non-directional relationship, use the following syntax. This will not specify a direction for the relationship, so relationships of any direction will be included in the result.

-- SELECT ... FROM ...
MATCH (source_twin_or_twin_collection)-[]-(target_twin_or_twin_collection)
-- WHERE ...

Tip

Non-directional queries require additional processing, which may result in increased latency and cost.

Examples

The first example shows a left-to-right directional traversal. This query finds twins Room and Factory where...

  • Room targets Factory (with any name of relationship)
  • Room has a temperature value that's greater than 50
  • Factory has a $dtId of 'ABC'
SELECT Room, Factory FROM DIGITALTWINS MATCH (Room)-[]->(Factory) 
WHERE Room.temperature > 50 AND Factory.$dtId = 'ABC'

Note

MATCH queries that contain $dtId filters on any twin other than the starting twin for the MATCH traversal may show empty results. This applies to factory.$dtId in the above example. For more information, see Limitations.

The following example shows a right-to-left directional traversal. This query looks similar to the one above, but the direction of the relationship between Room and Factory is reversed. This query finds twins Room and Factory where...

  • Factory targets Room (with any name of relationship)
  • Factory has a $dtId of 'ABC'
  • Room has a temperature value that's greater than 50
SELECT Room, Factory FROM DIGITALTWINS MATCH (Room)<-[]-(Factory) 
WHERE Factory.$dtId = 'ABC' AND Room.temperature > 50

The following example shows a non-directional traversal. This query finds twins Room and Factory where...

  • Room and Factory share any name of relationship, going in either direction
  • Factory has a $dtId of 'ABC'
  • Room has a humidity value that's greater than 70
SELECT Factory, Room FROM DIGITALTWINS MATCH (Factory)-[]-(Room) 
WHERE Factory.$dtId ='ABC'  AND Room.humidity > 70

Specify relationship name

Optionally, you can use the relationship condition in the MATCH clause to specify names for the relationships between the twins. You can specify a single name, or a list of possible names. The optional relationship name is included as part of the arrow syntax to specify relationship direction.

If you don't provide a relationship name, the query will include all relationship names by default.

Tip

Specifying relationship names in the query can improve performance and make results more predictable.

Syntax

Note

The examples in this section focus on relationship name. They all show non-directional relationships, they default to a single hop, and they don't assign query variables to the relationships. For instructions on how to do more with these other conditions, see Specify relationship direction, Specify number of hops, and Assign query variable to relationship. For information about how to use several of these together in the same query, see Combine MATCH operations.

Specify the name of a relationship to traverse in the MATCH clause within square brackets ([]), after a colon (:). This section shows the syntax of specifying named relationships.

For a single name, use the following syntax. The placeholder values that should be replaced with your values are twin_or_twin_collection_1, relationship_name, and twin_or_twin_collection_2.

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)-[:relationship_name]-(twin_or_twin_collection_2)
-- WHERE ...

For multiple possible names use the following syntax. The placeholder values that should be replaced with your values are twin_or_twin_collection_1, relationship_name_option_1, relationship_name_option_2, twin_or_twin_collection_2, and the note to continue the pattern as needed for the number of relationship names you want to enter.

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)-[:relationship_name_option_1|relationship_name_option_2|continue pattern as needed...]-(twin_or_twin_collection_2)
-- WHERE ...

Important

The colon (:) within the square brackets is a required part of the syntax for specifying a relationship name in a MATCH query. If you don't include the colon, your query doesn't specify a relationship name. Instead, you have a query that assigns a query variable to the relationship.

(Default) To leave name unspecified, leave the brackets empty of name information, like this:

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)-[]-(twin_or_twin_collection_2)
-- WHERE ...

Examples

The following example shows a single relationship name. This query finds twins Building and Sensor where...

  • Building has a 'contains' relationship to Sensor (going in either direction)
  • Building has a $dtId of 'Seattle21'
SELECT Building, Sensor FROM DIGITALTWINS   
MATCH (Building)-[:contains]-(Sensor)  
WHERE Building.$dtId = 'Seattle21'

The following example shows multiple possible relationship names. This query looks similar to the one above, but there are multiple possible relationship names that are included in the result. This query finds twins Building and Sensor where...

  • Building has either a 'contains' or 'isAssociatedWith' relationship to Sensor (going in either direction)
  • Building has a $dtId of 'Seattle21'
SELECT Building, Sensor FROM DIGITALTWINS   
MATCH (Building)-[:contains|isAssociatedWith]-(Sensor)  
WHERE building.$dtId = 'Seattle21'

The following example has no specified relationship name. As a result, relationships with any name will be included in the query result. This query finds twins Building and Sensor where...

  • Building has a relationship to Sensor with any name (and going in either direction)
  • Building has a $dtId of 'Seattle21'
SELECT Building, Sensor FROM DIGITALTWINS   
MATCH (Building-[]-(Sensor)  
WHERE Building.$dtId = 'Seattle21'

Specify number of hops

Optionally, you can use the relationship condition in the MATCH clause to specify the number of hops for the relationships between the twins. You can specify an exact number or a range. This optional value is included as part of the arrow syntax to specify relationship direction.

If you don't provide a number of hops, the query will default to one hop.

Important

If you specify a number of hops that is greater than one, you can't assign a query variable to the relationship. Only one of these conditions can be used within the same query.

Syntax

Note

The examples in this section focus on number of hops. They all show non-directional relationships without specifying names. For instructions on how to do more with these other conditions, see Specify relationship direction and Specify relationship name. For information about how to use several of these together in the same query, see Combine MATCH operations.

Specify the number of hops to traverse in the MATCH clause within the square brackets ([]).

To specify an exact number of hops, use the following syntax. The placeholder values that should be replaced with your values are twin_or_twin_collection_1, number_of_hops, and twin_or_twin_collection_2.

-- SELECT ... FROM ... 
MATCH (twin_or_twin_collection_1)-[*number_of_hops]-(twin_or_twin_collection_2)
-- WHERE ...

To specify a range of hops, use the following syntax. The placeholder values that should be replaced with your values are twin_or_twin_collection_1, starting_limit, ending_limit and twin_or_twin_collection_2. The starting limit isn't included in the range, while the ending limit is included.

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)-[*starting_limit..ending_limit]-(twin_or_twin_collection_2)
-- WHERE ...

You can also leave out the starting limit to indicate "anything up to" (and including) the ending limit. An ending limit must always be provided.

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)-[*..ending_limit]-(twin_or_twin_collection_2)
-- WHERE ...

(Default) To default to one hop, leave the brackets empty of hop information, like this:

-- SELECT ... FROM ... 
MATCH (twin_or_twin_collection_1)-[]-(twin_or_twin_collection_2)
-- WHERE ...

Examples

The following example specifies an exact number of hops. The query will only return relationships between twins Floor and Room that are exactly 3 hops.

SELECT Floor, Room FROM DIGITALTWINS 
MATCH (Floor)-[*3]-(Room)
WHERE Floor.$dtId = 'thermostat-15'

The following example specifies a range of hops. The query will return relationships between twins Floor and Room that are between 1 and 3 hops (meaning the number of hops is either 2 or 3).

SELECT Floor, Room FROM DIGITALTWINS 
MATCH (Floor)-[*1..3]-(Room)
WHERE Floor.$dtId = 'thermostat-15'

You can also show a range by providing only one boundary. In the following example, the query will return relationships between twins Floor and Room that are at most 2 hops (meaning the number of hops is either 1 or 2).

SELECT Floor, Room FROM DIGITALTWINS 
MATCH (Floor)-[*..2]-(Room)
WHERE Floor.$dtId = 'thermostat-15'

The following example has no specified number of hops, so will default to one hop between twins Floor and Room.

SELECT Floor, Room FROM DIGITALTWINS  
MATCH (Floor)-[]-(Room)
WHERE Floor.$dtId = 'thermostat-15'

Assign query variable to relationship (and specify relationship properties)

Optionally, you can assign a query variable to the relationship referenced in the MATCH clause, so that you can refer to it by name in the query text.

A useful result of doing this is the ability to filter on relationship properties in your WHERE clause.

Important

Assigning a query variable to the relationship is only supported when the query specifies a single hop. Within a query, you must choose between specifying a relationship variable and specifying a greater number of hops.

Syntax

Note

The examples in this section focus on a query variable for the relationship. They all show non-directional relationships without specifying names. For instructions on how to do more with these other conditions, see Specify relationship direction and Specify relationship name. For information about how to use several of these together in the same query, see Combine MATCH operations.

To assign a query variable to the relationship, put the variable name in the square brackets ([]). The placeholder values shown below that should be replaced with your values are twin_or_twin_collection_1, relationship_variable, and twin_or_twin_collection_2.

-- SELECT ... FROM ...   
MATCH (twin_or_twin_collection_1)-[relationship_variable]-(twin_or_twin_collection_2>) 
-- WHERE ...

Examples

The following example assigns a query variable 'Rel' to the relationship. Later, in the WHERE clause, it uses the variable to specify that the relationship Rel should have a name property with a value of 'child'.

SELECT Floor,Cafe, Rel DIGITALTWINS   
MATCH (Floor)-[Rel]-(Cafe)  
WHERE Floor.$dtId = 'thermostat-15' AND Rel.name = 'child'

Combine MATCH operations

You can combine multiple relationship conditions in the same query. You can also chain multiple relationship conditions to express bi-directional relationships or other larger combinations.

Syntax

In a single query, you can combine relationship direction, relationship name, and one of either number of hops or a query variable assignment.

The following syntax examples show how these attributes can be combined. You can also leave out any of the optional details shown in placeholders to omit that part of the condition.

To specify relationship direction, relationship name, and number of hops within a single query, use the following syntax within the relationship condition. The placeholder values that should be replaced with your values are twin_or_twin_collection_1 and twin_or_twin_collection_2, optional_left_angle_bracket and optional_right_angle_bracket, relationship_name(s), and number_of_hops.

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)optional_left_angle_bracket-[:relationship_name(s)*number_of_hops]-optional_right_angle_bracket(twin_or_twin_collection_2)
-- WHERE

To specify relationship direction, relationship name, and a query variable for the relationship within a single query, use the following syntax within the relationship condition. The placeholder values that should be replaced with your values are twin_or_twin_collection_1 and twin_or_twin_collection_2, optional_left_angle_bracket and optional_right_angle_bracket, relationship_variable, and relationship_name(s).

-- SELECT ... FROM ...
MATCH (twin_or_twin_collection_1)optional_left_angle_bracket-[relationship_variable:relationship_name(s)]-optional_right_angle_bracket(twin_or_twin_collection_2)
-- WHERE

Note

As per the options for specifying relationship direction, you must pick between a left angle bracket for a left-to-right relationship or a right angle bracket for a right-to-left relationship. You can't include both on the same arrow, but can represent bi-directional relationships by chaining.

You can chain multiple relationship conditions together, like this. The placeholder values that should be replaced with your values are twin_or_twin_collection_1, all instances of relationship_condition, and twin_or_twin_collection_2.

--Chained relationship conditions
-- SELECT ... FROM ... 
MATCH (twin_or_twin_collection_1)-[relationship_condition]-(twin_or_twin_collection_2)-[relationship_condition]-(twin_or_twin_collection_3)...
WHERE twin_or_twin_collection.$dtId = 'twin_ID'

Examples

Here's an example that combines relationship direction, relationship name, and number of hops. The following query finds twins Floor and Room where the relationship between Floor and Room meets these conditions:

  • the relationship is left-to-right, with Floor as the source and Room as the target
  • the relationship has a name of either 'contains' or 'isAssociatedWith'
  • the relationship has either 4 or 5 hops

The query also specifies that twin Floor has a $dtId of 'thermostat-15'.

SELECT Floor, Room FROM DIGITALTWINS    
MATCH (Floor)-[:contains|isAssociatedWith*3..5]->(Room) 
WHERE Floor.$dtId = 'thermostat-15'

Here's an example that combines relationship direction, relationship name, and a named query variable for the relationship. The following query finds twins Floor and Room where the relationship between Floor and Room is assigned to a query variable r and meets these conditions:

  • the relationship is left-to-right, with Floor as the source and Room as the target
  • the relationship has a name of either 'contains' or 'isAssociatedWith'
  • the relationship, which is given a query variable r, has a length property equal to 10

The query also specifies that twin Floor has a $dtId of 'thermostat-15'.

SELECT Floor, Room FROM DIGITALTWINS    
MATCH (Floor)-[r:contains|isAssociatedWith]->(Room) 
WHERE Floor.$dtId = 'thermostat-15' AND r.length = 10

The following example illustrates chained relationship conditions. The query finds twins Floor, Cafe, and Room, where...

  • the relationship between Floor and Room meets these conditions:
    • the relationship is left-to-right, with Floor as the source and Cafe as the target
    • the relationship has a name of either 'contains' or 'isAssociatedWith'
    • the relationship, which is given query variable r, has a length property equal to 10
  • the relationship between Cafe and Room meets these conditions:
    • the relationship is right-to-left, with Room as the source and Cafe as the target
    • the relationship has a name of either 'has' or 'includes'
    • the relationship has up to 3 (so 1, 2, or 3) hops

The query also specifies that twin Floor has a $dtId of 'thermostat-15' and twin Cafe has a temperature of 55.

SELECT Floor,Room, Cafe FROM DIGITALTWINS    
MATCH (Floor)-[r:contains|isAssociatedWith]->(Cafe)<-[has|includes*..3]-(Room)  
WHERE Floor.$dtId = 'thermostat-15'  AND r.length = 10 AND Cafe.temperature = 55

You can also use chained relationship conditions to express bi-directional relationships. The following query finds twins Floor, Room, and Building, where...

  • the relationship between Building and Floor meets these conditions:
    • the relationship is left-to-right, with Building as the source and Floor as the target
    • the relationship has a name of 'isAssociatedWith'
    • the relationship is given a query variable r1
  • the relationship between Floor and Room meets these conditions:
    • the relationship is right-to-left, with Room as the source and Floor as the target
    • the relationship has a name of 'isAssociatedWith'
    • the relationship is given a query variable r2

The query also specifies that twin Building has a $dtId of 'building-3' and Room has a temperature greater than 50.

SELECT Floor, Room, Building FROM DIGITALTWINS    
MATCH (Building)-[r1:isAssociatedWith]->(Floor)<-[r2:isAssociatedWith]-(Room)
WHERE Building.$dtId = 'building-3'  AND Room.temperature > 50

Limitations

The following limits apply to queries using MATCH:

  • Only one MATCH expression is supported per query statement.

  • $dtId is required in the WHERE clause.

  • Assigning a query variable to the relationship is only supported when the query specifies a single hop.

  • The maximum hops supported in a query is 10.

  • MATCH queries that contain $dtId filters on any twin other than the starting twin for the MATCH traversal may show empty results. For example, the following query is subject to this limitation:

    SELECT A, B, C FROM DIGITALTWINS 
    MATCH A-[contains]->B-[is_part_of]->C 
    WHERE B.$dtId = 'Device01'
    

    If your scenario requires you to use $dtId on other twins, consider using the JOIN clause instead.

  • MATCH queries that traverse the same twin multiple times may unexpectedly remove this twin from results.