Skip to main content
In this lesson you’ll provision a Cloud Spanner instance, create a database, define a parent table and an interleaved child table, add secondary indexes, insert sample data, and run queries in Spanner Studio (the in-console SQL editor). The instructions assume you are signed in to the Google Cloud Console. Getting started — create an instance
  1. Open Cloud Spanner from the GCP Console (left menu or search bar) and click Create Instance.
A screenshot of the Google Cloud Console's "Create an instance" page for Cloud Spanner. The form shows the instance name filled in as "spanner-demo-instance" with the Standard edition and a pricing estimate summary on the right.
Steps on the Create Instance form:
  1. Select the edition: choose Standard and click Continue.
  2. Provide an instance name, e.g., spanner-demo-instance (the instance ID is auto-filled).
  3. Choose a configuration: for this demo select Regional -> us-central1.
  4. For processing units, keep the default for the demo (note: production commonly starts at 100 PUs).
  5. Uncheck Enable default backup schedule for the demo (enable backups in production).
  6. Click Create. Instance creation typically takes 2–5 minutes.
Regional configurations provide 99.99% availability and are a good default for development and staging. For higher availability across regions, choose a multi-region configuration (99.999% SLA) but be aware of the increased cost.
After creation, go to All instances and refresh to confirm the instance status.
A screenshot of the Google Cloud Console showing the Cloud Spanner "Instances" page with a "Migrate to Cloud Spanner" banner. The instances table lists a "spanner-demo-instance" and shows details like region (us-central1), processing units, nodes, and storage utilization.
When the instance is READY, open it and click Create Database.
A Google Cloud Console screenshot showing the Cloud Spanner instance "spanner-demo-instance" overview. It displays instance details like region (us-central1), compute capacity (1000 PUs), and an instance summary with operations/throughput fields.
Database creation
  • Name the database, for example flightbookingdb.
  • Choose the SQL dialect: select Google Standard SQL (default). This dialect is based on SQL:2011 and is broadly compatible with PostgreSQL-style query syntax. A PostgreSQL dialect option exists if you need extra compatibility.
  • Click Create.
Open Spanner Studio (the SQL editor) for the database to create tables and run queries. Spanner Studio submits SQL to your Spanner instance and shows results, execution plans, and schema. Data types and best-fits Use Cloud Spanner-compatible types when designing the schema. The table below summarizes common choices used in this demo.
PurposeSpanner typeNotes
TextSTRING(MAX)Use for variable-length text.
IntegersINT6464-bit signed integers.
TimestampsTIMESTAMPUse with OPTIONS (allow_commit_timestamp = true) if you want commit timestamps.
Monetary / high-precision decimalsNUMERICFixed-point numeric type.
Create schema — parent and interleaved child table We’ll create a parent table named flights and an interleaved child table booking. Interleaving stores child rows physically with their parent, improving locality for parent-child queries. Create the parent table flights: 
CREATE TABLE flights (
  flight_id STRING(MAX) NOT NULL,
  flight_number STRING(MAX),
  airline STRING(MAX),
  origin STRING(MAX),
  destination STRING(MAX),
  departure_time TIMESTAMP,
  arrival_time TIMESTAMP,
  aircraft STRING(MAX),
  total_seats INT64,
  available_seats INT64
) PRIMARY KEY (flight_id);
Create the child table booking interleaved in flights. Note the child primary key begins with the parent’s primary key and that created_at allows commit timestamps: 
CREATE TABLE booking (
  flight_id STRING(MAX) NOT NULL,
  booking_id STRING(MAX) NOT NULL,
  passenger_name STRING(MAX),
  passenger_email STRING(MAX),
  seat_number STRING(MAX),
  booking_class STRING(MAX),
  booking_status STRING(MAX),
  booking_date TIMESTAMP,
  total_amount NUMERIC,
  created_at TIMESTAMP OPTIONS (allow_commit_timestamp = true)
) PRIMARY KEY (flight_id, booking_id),
  INTERLEAVE IN PARENT flights ON DELETE CASCADE;
Why this schema?
  • PRIMARY KEY (flight_id, booking_id) groups bookings under a flight and ensures uniqueness per booking.
  • INTERLEAVE IN PARENT places child rows alongside parent rows on storage, improving query performance for parent-child lookups and reducing cross-partition traffic.
  • ON DELETE CASCADE automatically removes child rows when the parent flight is deleted.
  • created_at ... allow_commit_timestamp = true enables use of PENDING_COMMIT_TIMESTAMP() in INSERTs so the timestamp is set to the actual transaction commit time.
Create secondary indexes Secondary indexes accelerate lookups by non-primary-key columns. Spanner maintains indexes automatically and chooses them in query plans as appropriate. 
CREATE INDEX bookings_by_email ON booking (passenger_email);
CREATE INDEX bookings_by_status ON booking (booking_status);
Verify the schema in Spanner Studio — you should see flights, booking, and the two indexes under the database schema. Insert sample data Insert sample flights into flights: 
INSERT INTO flights (
  flight_id, flight_number, airline, origin, destination,
  departure_time, arrival_time, aircraft, total_seats, available_seats
) VALUES
('FL1001', '1001', 'Acme Air', 'SFO', 'JFK', TIMESTAMP '2026-06-01T08:00:00Z', TIMESTAMP '2026-06-01T16:00:00Z', 'A320', 180, 180),
('FL1002', '1002', 'Acme Air', 'JFK', 'SFO', TIMESTAMP '2026-06-02T09:00:00Z', TIMESTAMP '2026-06-02T17:00:00Z', 'A320', 180, 180),
('FL1003', '1003', 'Contoso Airlines', 'SEA', 'LAX', TIMESTAMP '2026-06-03T07:30:00Z', TIMESTAMP '2026-06-03T09:30:00Z', '737', 160, 160);
Insert sample bookings. This example demonstrates PENDING_COMMIT_TIMESTAMP() for the created_at column: 
INSERT INTO booking (
  flight_id, booking_id, passenger_name, passenger_email,
  seat_number, booking_class, booking_status, booking_date,
  total_amount, created_at
) VALUES
('FL1001', 'BKG001', 'Alice Smith', 'alice@example.com', '12A', 'Economy', 'CONFIRMED', TIMESTAMP '2026-05-01T10:00:00Z', 199.99, PENDING_COMMIT_TIMESTAMP()),
('FL1001', 'BKG002', 'Bob Jones', 'bob@example.com', '12B', 'Economy', 'CONFIRMED', TIMESTAMP '2026-05-01T10:05:00Z', 199.99, PENDING_COMMIT_TIMESTAMP()),
('FL1002', 'BKG003', 'Carol Lee', 'carol@example.com', '3C', 'Business', 'CONFIRMED', TIMESTAMP '2026-05-02T11:00:00Z', 599.00, PENDING_COMMIT_TIMESTAMP());
PENDING_COMMIT_TIMESTAMP() sets the timestamp to the transaction commit time. For this to work the TIMESTAMP column must be declared with OPTIONS (allow_commit_timestamp = true). This is useful for audit trails and ensures all rows inserted in the same transaction share the exact same commit timestamp.
Run queries in Spanner Studio Examples you can run immediately in Spanner Studio:
  • Select all flights ordered by departure time:
SELECT *
FROM flights
ORDER BY departure_time;
  • Select bookings for a particular flight:
SELECT *
FROM booking
WHERE flight_id = 'FL1001';
  • Count total bookings:
SELECT COUNT(*) AS total_bookings
FROM booking;
Use the Explain / Execution details in Spanner Studio to inspect query plans, view scans, and see which indexes (if any) Spanner used. Design notes and best practices
  • Parent-child interleaving: child primary key must start with the parent primary key. This enables efficient locality and reduces cross-partition queries for parent-child access patterns.
  • Use interleaving for high-volume, locality-sensitive workloads (e.g., many bookings per flight). For small test datasets, the benefits may be less visible but become significant at scale.
  • Plan primary keys and access patterns carefully before deploying large distributed workloads; changing primary keys later is not straightforward.
  • Add secondary indexes selectively—indexes speed reads but add write overhead and storage costs. Choose indexes based on query patterns and cardinality.
  • Enable backups and monitoring in production. Use IAM roles to restrict who can modify schema and run administrative operations.
References and further reading That concludes this walkthrough on setting up Cloud Spanner, creating an interleaved schema, adding indexes, inserting sample data, and running queries in Spanner Studio.

Watch Video