Using SQL for Graph Queries

Joining Vertices and Edges in a Graph with SQL

1. Run the shell to submit the pre-edited demo GQL:

bin/gql_submit.sh --gql geaflow/geaflow-examples/gql/sql_join_to_graph_demo.sql

Here, sql_join_to_graph_demo.sql is an SQL Join query in a simulated streaming graph. Its key content is as follows:

USE GRAPH dy_modern;

select u.name, friend.name
from person u, knows e, person friend
where u.id = e.srcId and e.targetId = friend.id
;

This DSL reads node and edge data from the demo_job_data.txt resource file within the project to construct the graph.

Then, it performs a join query on nodes and edges of the graph dy_modern. The engine automatically translates the join semantics into a graph query.

2. Output Results

You can print the contents of the result file by running the following command:

cat /tmp/geaflow/sql_join_to_graph_demo_result/partition_0  

The query results are written to /tmp/geaflow/sql_join_to_graph_demo_result by default. Users can also customize the output path by modifying the geaflow.dsl.file.path parameter.

jim,jim
kate,kate
lily,lily
lucy,lucy
jim,jim
lucy,lucy
lucy,lucy
jack,jack

Using SQL to Perform Join and Aggregate Queries on Graph Vertices and Edges

1. Run the shell to submit the pre-edited demo GQL:

bin/gql_submit.sh --gql geaflow/geaflow-examples/gql/sql_join_to_graph_demo_02.sql

The key content of the sql_join_to_graph_demo_02.sql file is as follows:

USE GRAPH dy_modern;

SELECT    u.name,
          friend_num
FROM      person u,
          (
          SELECT    srcId AS pid,
                    COUNT(DISTINCT targetId) AS friend_num
          FROM      knows
          GROUP BY  srcId
          ) t_friend_num
WHERE     u.id = pid;

This DSL reads node and edge data from the demo_job_data.txt resource file within the project to construct the graph.

Then, it performs a join query on nodes and edges of the graph dy_modern. The engine automatically translates the join semantics into a graph query.

2. Output Results

You can print the contents of the result file by running the following command:

cat /tmp/geaflow/sql_join_to_graph_demo_02_result/partition_0  

The query results are written to /tmp/geaflow/sql_join_to_graph_demo_02_result by default. Users can also customize the output path by modifying the geaflow.dsl.file.path parameter.

The results show the number of friends recorded for each person in the graph:

lucy,2
jim,1
kate,1
lily,1
jack,1