Getting started with Spanner in C++

Prepare your local C++ environment

Clone the sample app repository to your local machine:

git clone https://github.com/googleapis/google-cloud-cpp $HOME/google-cloud-cpp

Install Bazel for Linux using these instructions.
Change to the directory that contains the Spanner sample code:
```
cd $HOME/google-cloud-cpp
```

Build the samples with this command:

bazel build //google/cloud/spanner/samples:samples

Set up authentication and authorization for the google-cloud-cpp project.
```
gcloud auth application-default login
```
Create an environment variable called PROJECT_ID. Replace [MY_PROJECT_ID] with your Google Cloud project ID. You can find this ID in your project's Welcome page.
```
export PROJECT_ID=[MY_PROJECT_ID]
```

Create an instance

When you first use Spanner, you must create an instance, which is an allocation of resources that are used by Spanner databases. When you create an instance, you choose an instance configuration, which determines where your data is stored, and also the number of nodes to use, which determines the amount of serving and storage resources in your instance.

See Create an instance to learn how to create a Spanner instance using any of the following methods. You can name your instance test-instance to use it with other topics in this document that reference an instance named test-instance.

The Google Cloud CLI
The Google Cloud console
A client library (C++, C#, Go, Java, Node.js, PHP, Python, or Ruby)

Look through sample files

The samples repository contains a sample that shows how to use Spanner with C++.

Take a look through the google/cloud/spanner/samples/samples.cc file, which shows how to create a database and modify a database schema. The data uses the example schema shown in the Schema and data model page.

Create a database

GoogleSQL

bazel run //google/cloud/spanner/samples:samples -- \
      create-database PROJECT_ID test-instance example-db

PostgreSQL

bazel run //google/cloud/spanner/samples:postgresql_samples -- \
      create-database PROJECT_ID test-instance example-db

bazel run //google/cloud/spanner/samples:postgresql_samples -- \
      interleaved-table PROJECT_ID test-instance example-db

You should see:

Created database [projects/${PROJECT_ID}/instances/test-instance/databases/example-db]

The following code creates a database and two tables in the database.

GoogleSQL

PostgreSQL

In the PostgreSQL dialect, the database needs to be created before submitting a DDL request to create a table.

The following example creates a database:

The following example creates the two tables in the database:

The next step is to write data to your database.

Create a database client

Before you can do reads or writes, you must create a Client:

A Client lets you read, write, query, and execute transactions on a Spanner database. Typically you create a Client when your application starts up, then you re-use that Client to read, write, and execute transactions. Each client uses resources in Spanner. The destructor of Client automatically cleans up the Client resources, including network connections.

Read more about Client in the Google Cloud Spanner C++ Reference.

Write data with DML

You can insert data using Data Manipulation Language (DML) in a read-write transaction.

You use the Client::ExecuteDml() function to execute a DML statement.

Run the sample using the getting-started-insert argument.

bazel run //google/cloud/spanner/samples:samples -- \
    getting-started-insert PROJECT_ID test-instance example-db

You should see:

Insert was successful [spanner_dml_getting_started_insert]

Write data with mutations

You can also insert data using mutations.

You write data using a Client object. The Client::Commit() function creates and commits a transaction for writes that execute atomically at a single logical point in time across columns, rows, and tables in a database.

This code shows how to write the data using mutations:

Run the sample using the insert-data argument.

bazel run //google/cloud/spanner/samples:samples -- \
    insert-data PROJECT_ID test-instance example-db

You should see:

Insert was successful [spanner_insert_data]

Query data using SQL

Spanner supports a SQL interface for reading data, which you can access on the command line using the Google Cloud CLI or programmatically using the Spanner client library for C++.

On the command line

Execute the following SQL statement to read the values of all columns from the Albums table:

gcloud spanner databases execute-sql example-db --instance=test-instance \
    --sql='SELECT SingerId, AlbumId, AlbumTitle FROM Albums'

The result shows:

SingerId AlbumId AlbumTitle
1        1       Total Junk
1        2       Go, Go, Go
2        1       Green
2        2       Forever Hold Your Peace
2        3       Terrified

Use the Spanner client library for C++

In addition to executing a SQL statement on the command line, you can issue the same SQL statement programmatically using the Spanner client library for C++.

You use the Client::ExecuteQuery() function to run the SQL query. Here's how to issue the query and access the data:

Run the sample using the query_data argument.

bazel run //google/cloud/spanner/samples:samples -- \
    query-data PROJECT_ID test-instance example-db

You should see the following result:

SingerId: 1     LastName: Richards
SingerId: 2     LastName: Smith
SingerId: 3     LastName: Trentor
SingerId: 4     LastName: Martin
SingerId: 5     LastName: Lomond
SingerId: 12    LastName: Garcia
SingerId: 13    LastName: Morales
SingerId: 14    LastName: Long
SingerId: 15    LastName: Shaw

Query using a SQL parameter

If your application has a frequently executed query, you can improve its performance by parameterizing it. The resulting parametric query can be cached and reused, which reduces compilation costs. For more information, see Use query parameters to speed up frequently executed queries.

Here is an example of using a parameter in the WHERE clause to query records containing a specific value for LastName.

GoogleSQL

PostgreSQL

Run the sample using the query-with-parameter command.

bazel run //google/cloud/spanner/samples:samples -- \
    query-with-parameter PROJECT_ID test-instance example-db

You should see the following result:

SingerId: 12    FirstName: Melissa      LastName: Garcia

Read data using the read API

In addition to Spanner's SQL interface, Spanner also supports a read interface.

You use the Client::Read() function to read rows from the database. Use a KeySet object to define a collection of keys and key ranges to read.

Here's how to read the data:

Run the sample using the read-data argument.

bazel run //google/cloud/spanner/samples:samples -- \
    read-data PROJECT_ID test-instance example-db

You should see output similar to:

SingerId: 1, AlbumId: 1, AlbumTitle: Total Junk
SingerId: 1, AlbumId: 2, AlbumTitle: Go, Go, Go
SingerId: 2, AlbumId: 1, AlbumTitle: Green
SingerId: 2, AlbumId: 2, AlbumTitle: Forever Hold Your Peace
SingerId: 2, AlbumId: 3, AlbumTitle: Terrified

Update the database schema

Assume you need to add a new column called MarketingBudget to the Albums table. Adding a new column to an existing table requires an update to your database schema. Spanner supports schema updates to a database while the database continues to serve traffic. Schema updates don't require taking the database offline and they don't lock entire tables or columns; you can continue writing data to the database during the schema update. Read more about supported schema updates and schema change performance in Make schema updates.

Add a column

You can add a column on the command line using the Google Cloud CLI or programmatically using the Spanner client library for C++.

On the command line

Use the following ALTER TABLE command to add the new column to the table:

GoogleSQL

gcloud spanner databases ddl update example-db --instance=test-instance \
    --ddl='ALTER TABLE Albums ADD COLUMN MarketingBudget INT64'

PostgreSQL

gcloud spanner databases ddl update example-db --instance=test-instance \
    --ddl='ALTER TABLE Albums ADD COLUMN MarketingBudget BIGINT'

You should see:

Schema updating...done.

Use the Spanner client library for C++

Use the DatabaseAdminClient::UpdateDatabase() function to modify the schema.

GoogleSQL

PostgreSQL

Run the sample using the add-column command.

bazel run //google/cloud/spanner/samples:samples -- \
    add-column PROJECT_ID test-instance example-db

You should see:

Added MarketingBudget column

Write data to the new column

The following code writes data to the new column. It sets MarketingBudget to 100000 for the row keyed by Albums(1, 1) and to 500000 for the row keyed by Albums(2, 2).

Run the sample using the update-data argument.

bazel run //google/cloud/spanner/samples:samples -- \
    update-data PROJECT_ID test-instance example-db

You can also execute a SQL query or a read call to fetch the values that you just wrote.

Here's the code to execute the query:

To execute this query, run the sample using the query-new-column argument.

bazel run //google/cloud/spanner/samples:samples -- \
    query-new-column PROJECT_ID test-instance example-db

You should see:

SingerId: 1 AlbumId: 1  MarketingBudget: 100000
SingerId: 1 AlbumId: 2  MarketingBudget: NULL
SingerId: 2 AlbumId: 1  MarketingBudget: NULL
SingerId: 2 AlbumId: 2  MarketingBudget: 500000
SingerId: 2 AlbumId: 3  MarketingBudget: NULL

Update data

You can update data using DML in a read-write transaction.

You use the Client::ExecuteDml() function to execute a DML statement.

GoogleSQL

PostgreSQL

Run the sample using the getting-started-update argument.

bazel run //google/cloud/spanner/samples:samples -- \
    getting-started-update PROJECT_ID test-instance example-db

You should see:

Update was successful [spanner_dml_getting_started_update]

Use a secondary index

Suppose you wanted to fetch all rows of Albums that have AlbumTitle values in a certain range. You could read all values from the AlbumTitle column using a SQL statement or a read call, and then discard the rows that don't meet the criteria, but doing this full table scan is expensive, especially for tables with a lot of rows. Instead you can speed up the retrieval of rows when searching by non-primary key columns by creating a secondary index on the table.

Adding a secondary index to an existing table requires a schema update. Like other schema updates, Spanner supports adding an index while the database continues to serve traffic. Spanner automatically backfills the index with your existing data. Backfills might take a few minutes to complete, but you don't need to take the database offline or avoid writing to the indexed table during this process. For more details, see Add a secondary index.

After you add a secondary index, Spanner automatically uses it for SQL queries that are likely to run faster with the index. If you use the read interface, you must specify the index that you want to use.

Add a secondary index

You can add an index on the command line using the gcloud CLI or programmatically using the Spanner client library for C++.

On the command line

Use the following CREATE INDEX command to add an index to the database:

gcloud spanner databases ddl update example-db --instance=test-instance \
    --ddl='CREATE INDEX AlbumsByAlbumTitle ON Albums(AlbumTitle)'

You should see:

Schema updating...done.

Using the Spanner client library for C++

You use the DatabaseAdminClient::UpdateDatabase() function to add an index:

Run the sample using the add-index argument.

bazel run //google/cloud/spanner/samples:samples -- \
    add-index PROJECT_ID test-instance example-db

Adding an index can take a few minutes. After the index is added, you should see output similar to this:

`AlbumsByAlbumTitle` Index successfully added, new DDL:
database: "projects/PROJECT_ID/instances/test-instance/databases/example-db"
statements: "CREATE INDEX AlbumsByAlbumTitle ON Albums(AlbumTitle)"
commit_timestamps {
  seconds: 1581011550
  nanos: 531102000
}

Read using the index

For SQL queries, Spanner automatically uses an appropriate index. In the read interface, you must specify the index in your request.

To use the index in the read interface, use the Client::Read() function, which reads zero or more rows from a database using an index.

The following code fetches all AlbumId, and AlbumTitle columns from the AlbumsByAlbumTitle index.

Run the sample using the read-data-with-index argument.

bazel run //google/cloud/spanner/samples:samples -- \
    read-data-with-index PROJECT_ID test-instance example-db

You should see:

AlbumId: 2  AlbumTitle: Forever Hold Your Peace
AlbumId: 2  AlbumTitle: Go, Go, Go
AlbumId: 1  AlbumTitle: Green
AlbumId: 3  AlbumTitle: Terrified
AlbumId: 1  AlbumTitle: Total Junk

Add an index for index-only reads

You might have noticed that the previous read example doesn't include reading the MarketingBudget column. This is because Spanner's read interface doesn't support the ability to join an index with a data table to look up values that are not stored in the index.

Create an alternate definition of AlbumsByAlbumTitle that stores a copy of MarketingBudget in the index.

On the command line

GoogleSQL

gcloud spanner databases ddl update example-db --instance=test-instance \
    --ddl='CREATE INDEX AlbumsByAlbumTitle2 ON Albums(AlbumTitle) STORING (MarketingBudget)

PostgreSQL

gcloud spanner databases ddl update example-db --instance=test-instance \
    --ddl='CREATE INDEX AlbumsByAlbumTitle2 ON Albums(AlbumTitle) INCLUDE (MarketingBudget)

Adding an index can take a few minutes. After the index is added, you should see:

Schema updating...done.

Using the Spanner client library for C++

You use the DatabaseAdminClient::UpdateDatabase() function to add an index with a STORING clause for :

Run the sample using the add-storing-index argument.

bazel run //google/cloud/spanner/samples:samples -- \
    add-storing-index PROJECT_ID test-instance example-db

You should see output similar to this:

`AlbumsByAlbumTitle2` Index successfully added, new DDL:
database: "projects/PROJECT_ID/instances/test-instance/databases/example-db"
statements: "CREATE INDEX AlbumsByAlbumTitle2 ON Albums(AlbumTitle) STORING (MarketingBudget)"
commit_timestamps {
  seconds: 1581012328
  nanos: 416682000
}

Now you can execute a read that fetches all AlbumId, AlbumTitle, and MarketingBudget columns from the AlbumsByAlbumTitle2 index:

Read data using the storing index you created by executing a query that explicitly specifies the index:

Run the sample using the read-data-with-storing-index argument.

bazel run //google/cloud/spanner/samples:samples -- \
    read-data-with-storing-index PROJECT_ID test-instance example-db

You should see output similar to:

AlbumId: 2  AlbumTitle: Forever Hold Your Peace MarketingBudget: 520000
AlbumId: 2  AlbumTitle: Go, Go, Go  MarketingBudget: NULL
AlbumId: 1  AlbumTitle: Green   MarketingBudget: NULL
AlbumId: 3  AlbumTitle: Terrified   MarketingBudget: NULL
AlbumId: 1  AlbumTitle: Total Junk  MarketingBudget: 80000

Retrieve data using read-only transactions

Suppose you want to execute more than one read at the same timestamp. Read-only transactions observe a consistent prefix of the transaction commit history, so your application always gets consistent data. The Transaction type is used to represent all kinds of transactions. Use the MakeReadOnlyTransaction() factory function to create a read-only transaction.

The following shows how to run a query and perform a read in the same read-only transaction:

Run the sample using the read-only-transaction argument.

bazel run //google/cloud/spanner/samples:samples -- \
    read-only-transaction PROJECT_ID test-instance example-db

You should see output similar to:

Read 1 results
SingerId: 2 AlbumId: 2 AlbumTitle: Forever Hold Your Peace
SingerId: 1 AlbumId: 2 AlbumTitle: Go, Go, Go
SingerId: 2 AlbumId: 1 AlbumTitle: Green
SingerId: 2 AlbumId: 3 AlbumTitle: Terrified
SingerId: 1 AlbumId: 1 AlbumTitle: Total Junk
Read 2 results
SingerId: 2 AlbumId: 2 AlbumTitle: Forever Hold Your Peace
SingerId: 1 AlbumId: 2 AlbumTitle: Go, Go, Go
SingerId: 2 AlbumId: 1 AlbumTitle: Green
SingerId: 2 AlbumId: 3 AlbumTitle: Terrified
SingerId: 1 AlbumId: 1 AlbumTitle: Total Junk

Cleanup

To avoid incurring additional charges to your Cloud Billing account for the resources used in this tutorial, drop the database and delete the instance that you created.

Delete the database

If you delete an instance, all databases within it are automatically deleted. This step shows how to delete a database without deleting an instance (you would still incur charges for the instance).

On the command line

gcloud spanner databases delete example-db --instance=test-instance

Using the Google Cloud console

Go to the Spanner Instances page in the Google Cloud console.

Go to the Instances page
Click the instance.
Click the database that you want to delete.
In the Database details page, click Delete.
Confirm that you want to delete the database and click Delete.

Delete the instance

Deleting an instance automatically drops all databases created in that instance.

On the command line

gcloud spanner instances delete test-instance

Using the Google Cloud console

Go to the Spanner Instances page in the Google Cloud console.

Go to the Instances page
Click your instance.
Click Delete.
Confirm that you want to delete the instance and click Delete.