Query Operations

A GSQL query is a sequence of retrieval and computation statements executed as a single transaction.

1. Creating a Query

Use CREATE QUERY to define a procedure. Queries can accept parameters and return values.

gsqlterminal
Copy
CREATE QUERY my_query (STRING name, INT age) FOR GRAPH MyGraph {
  # Body of the query
  PRINT "Hello " + name;
}

Key Keywords:

• OR REPLACE: Updates an existing query definition.
• DISTRIBUTED: Optimizes the query for execution across a cluster.
• RETURNS: Specifies the data type returned when used as a subquery.

2. AD-HOC vs. Compiled Execution

TigerGraph offers two ways to run queries:

Interpreted Mode (INTERPRET)

Ideal for development and ad-hoc analysis. It runs immediately without compilation.

bashterminal
Copy
INTERPRET QUERY () FOR GRAPH MyGraph { PRINT "Quick check"; }

Installed Mode (INSTALL + RUN)

Required for production. Compilation optimizes the query for maximum performance and creates a REST endpoint.

bashterminal
Copy
INSTALL QUERY my_query
RUN QUERY my_query("Alice", 30)

3. Dynamic Querying

You can parameterize not just data, but the schema itself. This allows a single query to work on different vertex/edge types or attributes.

gsqlterminal
Copy
CREATE QUERY dynamic_search (STRING v_type, STRING attr_name) {
  Start = {v_type};
  V = SELECT s FROM Start:s
      POST-ACCUM s.setAttr(attr_name, "Updated");
}

4. Query Lifecycle Commands

• SHOW QUERY queryName: View the source code.
• DROP QUERY queryName: Remove the query from the catalog.
• INSTALL QUERY ALL: Compiles all uninstalled queries in one batch.