The interaction with the database normally starts with a call to connect(), which establishes a new database session. This results in getting a connection Promise. The connection instance provides methods to run queries.

Copy
const edgedb = require("edgedb");

async function main() {
  // Establish a connection to an existing database
  // named "test" as an "edgedb" user.
  const conn = await edgedb.connect("edgedb://edgedb@localhost/test");

  try {
    // Create a User object type.
    await conn.execute(`
      CREATE TYPE User {
          CREATE REQUIRED PROPERTY name -> str;
          CREATE PROPERTY dob -> cal::local_date;
      }
    `);

    // Insert a new User object.
    await conn.query(
      `
      INSERT User {
          name := <str>$name,
          dob := <cal::local_date>$dob
      }
      `,
      {
        name: "Bob",
        dob: new edgedb.LocalDate(1984, 3, 1)
      }
    );

    // Select User objects.
    const userSet = await conn.query(
      "SELECT User {name, dob} FILTER .name = <str>$name",
      { name: "Bob" }
    );
    userSet; // [{ name: 'Bob', dob: edgedb.LocalDate(1984, 3, 1) }]

  } finally {
    conn.close()
  }
}

main();

For details on usage with TypeScript, go to TypeScript.

edgedb automatically converts EdgeDB types to the corresponding JavaScript types and vice versa. See Datatypes for details.

For server-side applications that handle frequent requests and need database connections for short periods of time, it is recommended to use a connection pool. The edgedb-js API provides an implementation of such pool.

To create a connection pool, use the createPool() method. The resulting edgedb.Pool object can be used to maintain a certain number of open connections and borrow them when needed.

Copy
const edgedb = require("edgedb");

async function main() {
  // Create a connection pool to an existing database
  const pool = await edgedb.createPool("edgedb://edgedb@localhost/test");

  try {
    // Create a User object type.
    await pool.execute(`
      CREATE TYPE User {
          CREATE REQUIRED PROPERTY name -> str;
          CREATE PROPERTY dob -> cal::local_date;
      }
    `);

    // Insert a new User object.
    await pool.query(
      `
       INSERT User {
         name := <str>$name,
         dob := <cal::local_date>$dob
       }
      `,
      {
        name: "Bob",
        dob: new edgedb.LocalDate(1984, 3, 1)
      }
    );

    // Select User objects.
    let userSet = await pool.query(
      "SELECT User {name, dob} FILTER .name = <str>$name",
      { name: "Bob" }
    );
    userSet; // [{ name: 'Bob', dob: edgedb.LocalDate(1984, 3, 1) }]

  } finally {
    await pool.close();
  }
}

main();

See Pool API documentation for more information.

The most robust way to execute transactional code is to use the retryingTransaction() API:

Copy
await pool.retryingTransaction(tx => {
  await tx.execute("INSERT User {name := 'Don'}");
});

Note that we execute queries on the tx object in the above example, rather than on the original connection pool pool object.

The retryingTransaction() API guarantees that:

  1. Transactions are executed atomically;

  2. If a transaction is failed for any of the number of transient errors (i.e. a network failure or a concurrent update error), the transaction would be retried;

  3. If any other, non-retryable exception occurs, the transaction is rolled back, and the exception is propagated, immediately aborting the retryingTransaction() block.

The key implication of retrying transactions is that the entire nested code block can be re-run, including any non-querying JavaScript code. Here is an example:

Copy
pool.retryingTransaction(tx => {
    let user = await tx.querySingle(
        "SELECT User { email } FILTER .login = <str>$login",
        login=login,
    )
    let query = await fetch(
        'https://service.local/email_info', {
            body: JSON.stringify({email: user.email})
            headers: { 'Content-Type': 'application/json' },
        },
    )
    let data = await query.json()
    user = await tx.querySingle('''
            UPDATE User FILTER .login = <str>$login
            SET { email_info := <json>$data}
        ''',
        login=login,
        data=data,
    )
})

In the above example, the execution of the HTTP request would be retried too. The core of the issue is that whenever transaction is interrupted user might have the email changed (as the result of concurrent transaction), so we have to redo all the work done.

Generally it’s recommended to not execute any long running code within the transaction unless absolutely necessary.

Transactions allocate expensive server resources and having too many concurrently running long-running transactions will negatively impact the performance of the DB server.

See also:

Light
Dark
System