This page discusses schemas and introduces interleaved tables, which can improve query performance when querying tables in a parent-child relationship.
Cloud Spanner databases contain one or more tables. Tables are structured as rows and columns. One or more of the columns are defined as the table's primary key, which uniquely identifies each row. Primary keys are always indexed for quick row lookup, and you can define secondary indexes on one or more columns. If you want to update or delete existing rows in a table, then the table must have a primary key. A table with no primary key columns can have only one row. Only GoogleSQL-dialect databases can have tables without a primary key.
Data in Spanner is strongly typed. You must define a schema for each database, and that schema must specify the data type of each column of each table. Data types include scalar and complex types, which are described in Data types in GoogleSQL and PostgreSQL data types.
Parent-child table relationships
There are two ways to define parent-child relationships in Spanner: table interleaving and foreign keys.
Spanner's table interleaving is a good choice for many parent-child
relationships. With interleaving, Spanner physically co-locates child
rows with parent rows in storage. Co-location can
significantly improve performance. For example, if you have a Customers table
and an Invoices table, and your application frequently fetches all the
invoices for a customer, you can define Invoices as an interleaved child
table of Customers. In doing so, you're declaring a data locality
relationship between two independent tables. You're telling Spanner
to store one or more rows of Invoices with one Customers row.
You associate a child table with a parent table by using DDL that declares the child table as interleaved in the parent, and by including the parent table primary key as the first part of the child table composite primary key. For more information about interleaving, see Create interleaved tables later in this topic.
Foreign keys are a more general parent-child solution and address additional use cases. They are not limited to primary key columns, and tables can have multiple foreign key relationships, both as a parent in some relationships and a child in others. However, a foreign key relationship does not imply co-location of the tables in the storage layer.
Google recommends that you choose to represent parent-child relationships either as interleaved tables or as foreign keys, but not both. For more information on foreign keys and their comparison to interleaved tables, see Foreign keys overview.
Choosing a primary key
Often your application already has a field that's a natural
fit for use as the primary key. For example, for a Customers table,
there might be an application-supplied CustomerId that serves well as
the primary key. In other cases, you may need to generate a primary key when
inserting the row. This would typically be a unique integer value with no
business significance (a surrogate primary key).
In all cases, you should be careful not to create hotspots with the choice of your primary key. For example, if you insert records with a monotonically increasing integer as the key, you'll always insert at the end of your key space. This is undesirable because Spanner divides data among servers by key ranges, which means your inserts will be directed at a single server, creating a hotspot. There are techniques that can spread the load across multiple servers and avoid hotspots:
- Hash the key and store it in a column. Use the hash column (or the hash column and the unique key columns together) as the primary key.
- Swap the order of the columns in the primary key.
- Use a Universally Unique Identifier (UUID). Version 4 UUID is recommended, because it uses random values in the high-order bits. Don't use a UUID algorithm (such as version 1 UUID) that stores the timestamp in the high order bits.
- Bit-reverse sequential values.
Adding secondary indexes based on primary keys
In certain circumstances, your database usage can benefit from adding secondary indexes based on primary keys. This is particularly true if you frequently run queries that require reverse-order scans of a table's primary key.
Primary keys in interleaved tables
For interleaving, every table must have a primary key. If you declare a table to be an interleaved child of another table, the table must have a composite primary key that includes all of the components of the parent's primary key, in the same order, and, typically, one or more additional child table columns.
Spanner stores rows in sorted order by primary key values, with child rows inserted between parent rows. See an illustration of interleaved rows in Create interleaved tables later in this topic.
In summary, Spanner can physically co-locate rows of related tables. The schema examples show what this physical layout looks like.
Database splits
You can define hierarchies of interleaved parent-child relationships up to seven layers deep, which means that you can co-locate rows of seven independent tables. If the size of the data in your tables is small, a single Spanner server can probably handle your database. But what happens when your related tables grow and start reaching the resource limits of an individual server? Spanner is a distributed database, which means that as your database grows, Spanner divides your data into chunks called "splits." Individual splits can move independently from each other and get assigned to different servers, which can be in different physical locations. A split holds a range of contiguous rows. The start and end keys of this range are called "split boundaries". Spanner automatically adds and removes split boundaries based on size and load, which changes the number of splits in the database.
Load-based splitting
As an example of how Spanner performs load-based splitting to mitigate read hotspots, suppose your database contains a table with 10 rows that are read more frequently than all of the other rows in the table. Spanner can add split boundaries between each of those 10 rows so that they're each handled by a different server, rather than allowing all the reads of those rows to consume the resources of a single server.
As a general rule, if you follow best practices for schema design, Spanner can mitigate hotspots such that the read throughput should improve every few minutes until you saturate the resources in your instance or run into cases where no new split boundaries can be added (because you have a split that covers just a single row with no interleaved children).
Schema examples
The schema examples below show how to create parent and child tables with and without interleaving, and illustrate the corresponding physical layouts of data.
Create a parent table
Suppose you're creating a music application and you need a simple table that stores rows of singer data:
Note that the table contains one primary key column, SingerId, which appears
to the left of the bolded line, and that tables are organized by rows
and columns.
You can define the table with a Spanner schema like this:
GoogleSQL
CREATE TABLE Singers ( SingerId INT64 NOT NULL, FirstName STRING(1024), LastName STRING(1024), SingerInfo BYTES(MAX), ) PRIMARY KEY (SingerId);
PostgreSQL
CREATE TABLE singers ( singer_id BIGINT PRIMARY KEY, first_name VARCHAR(1024), last_name VARCHAR(1024), singer_info BYTEA );
Note the following about the example schema:
Singersis a table at the root of the database hierarchy (because it's not defined as an interleaved child of another table).- For GoogleSQL-dialect databases, primary key columns are usually annotated with
NOT NULL(though you can omit this annotation if you want to allowNULLvalues in key columns. For more information, see Key Columns). - Columns that are not included in the primary key are called non-key columns,
and they can have an optional
NOT NULLannotation. - Columns that use the
STRINGorBYTEStype in GoogleSQL must be defined with a length, which represents the maximum number of Unicode characters that can be stored in the field. The length specification is optional for the PostgreSQLvarcharandcharacter varyingtypes. For more information, see Scalar Data Types for GoogleSQL-dialect databases and PostgreSQL data types for PostgreSQL-dialect databases.
What does the physical layout of the rows in the Singers table look like? The
following diagram shows rows of the Singers table stored by primary key
("Singers(1)", then "Singers(2)", and so
on, where the number in parentheses is the primary key value.
The preceding diagram illustrates an example split boundary between the rows keyed
by Singers(3) and Singers(4), with the data from the resulting splits
assigned to different servers. As this table grows, it's
possible for rows of Singers data to be stored in different locations.
Create parent and child tables
Assume that you now want to add some basic data about each singer's albums to the music application.
Note that the primary key of Albums is composed of two columns: SingerId and
AlbumId, to associate each album with its singer. The following example schema
defines both the Albums and Singers tables at the root of the database
hierarchy, which makes them sibling tables.
-- Schema hierarchy: -- + Singers (sibling table of Albums) -- + Albums (sibling table of Singers)
GoogleSQL
CREATE TABLE Singers ( SingerId INT64 NOT NULL, FirstName STRING(1024), LastName STRING(1024), SingerInfo BYTES(MAX), ) PRIMARY KEY (SingerId); CREATE TABLE Albums ( SingerId INT64 NOT NULL, AlbumId INT64 NOT NULL, AlbumTitle STRING(MAX), ) PRIMARY KEY (SingerId, AlbumId);
PostgreSQL
CREATE TABLE singers ( singer_id BIGINT PRIMARY KEY, first_name VARCHAR(1024), last_name VARCHAR(1024), singer_info BYTEA ); CREATE TABLE albums ( singer_id BIGINT, album_id BIGINT, album_title VARCHAR, PRIMARY KEY (singer_id, album_id) );
The physical layout of the rows of Singers and Albums looks like the following
diagram, with rows of the Albums table stored by contiguous primary key, then
rows of Singers stored by contiguous primary key:
One important note about the schema is that Spanner assumes no
data locality relationships between the Singers and Albums tables, because
they are top-level tables. As the database grows, Spanner can add split
boundaries between any of the rows. This means the rows of the
Albums table could end up in a different split from the rows of the Singers
table, and the two splits could move independently from each other.
Depending on your application's needs, it might be fine to allow Albums data
to be located on different splits from Singers data. However, this might incur
a performance penalty due to the need to coordinate reads and updates across
distinct resources. If your
application frequently needs to retrieve information about all the albums for a
particular singer, then you should create Albums as an interleaved child table
of Singers, which co-locates rows from the two tables along the primary key
dimension. The next example explains this in more detail.
Create interleaved tables
An interleaved table is a table that you declare to be an interleaved child of another table because you want the rows of the child table to be physically stored with the associated parent row. As mentioned earlier, the parent table primary key must be the first part of the child table composite primary key.
As you're designing your music application, suppose you realize that the app
needs to frequently access rows from the Albums table when it accesses a
Singers row. For example, when you access the row Singers(1), you also need
to access the rows Albums(1, 1) and Albums(1, 2). In this case, Singers
and Albums need to have a strong data locality relationship.
You can declare this data locality relationship by creating Albums as an
interleaved child table of Singers.
-- Schema hierarchy: -- + Singers -- + Albums (interleaved table, child table of Singers)
The bolded line in the following schema shows how to create Albums as an
interleaved table of Singers.
GoogleSQL
CREATE TABLE Singers ( SingerId INT64 NOT NULL, FirstName STRING(1024), LastName STRING(1024), SingerInfo BYTES(MAX), ) PRIMARY KEY (SingerId); CREATE TABLE Albums ( SingerId INT64 NOT NULL, AlbumId INT64 NOT NULL, AlbumTitle STRING(MAX), ) PRIMARY KEY (SingerId, AlbumId), INTERLEAVE IN PARENT Singers ON DELETE CASCADE;
PostgreSQL
CREATE TABLE singers ( singer_id BIGINT PRIMARY KEY, first_name VARCHAR(1024), last_name VARCHAR(1024), singer_info BYTEA ); CREATE TABLE albums ( singer_id BIGINT, album_id BIGINT, album_title VARCHAR, PRIMARY KEY (singer_id, album_id) ) INTERLEAVE IN PARENT singers ON DELETE CASCADE;
Notes about this schema:
SingerId, which is the first part of the primary key of the child tableAlbums, is also the primary key of its parent tableSingers.- The
ON DELETE CASCADEannotation signifies that when a row from the parent table is deleted, its child rows are automatically deleted as well. If a child table doesn't have this annotation, or the annotation isON DELETE NO ACTION, then you must delete the child rows before you can delete the parent row. - Interleaved rows are ordered first by rows of the parent table, then by contiguous rows of the child table that share the parent's primary key. For example, "Singers(1)", then "Albums(1, 1)", then "Albums(1, 2)", and so on.
- The data locality relationship of each singer and their album data is
preserved if this database splits, provided that the size of a
Singersrow and all itsAlbumsrows stays below the split size limit and that there is no hotspot in any of theseAlbumsrows. - The parent row must exist before you can insert child rows. The parent row can either already exist in the database or can be inserted before the insertion of the child rows in the same transaction.
Create a hierarchy of interleaved tables
The parent-child relationship between Singers and Albums can be extended to
more descendant tables. For example, you could create an interleaved table
called Songs as a child of Albums to store the track list of each album:
Songs must have a primary key that includes all the primary keys of the
tables above it in the hierarchy, that is, SingerId and AlbumId.
-- Schema hierarchy: -- + Singers -- + Albums (interleaved table, child table of Singers) -- + Songs (interleaved table, child table of Albums)
GoogleSQL
CREATE TABLE Singers ( SingerId INT64 NOT NULL, FirstName STRING(1024), LastName STRING(1024), SingerInfo BYTES(MAX), ) PRIMARY KEY (SingerId); CREATE TABLE Albums ( SingerId INT64 NOT NULL, AlbumId INT64 NOT NULL, AlbumTitle STRING(MAX), ) PRIMARY KEY (SingerId, AlbumId), INTERLEAVE IN PARENT Singers ON DELETE CASCADE; CREATE TABLE Songs ( SingerId INT64 NOT NULL, AlbumId INT64 NOT NULL, TrackId INT64 NOT NULL, SongName STRING(MAX), ) PRIMARY KEY (SingerId, AlbumId, TrackId), INTERLEAVE IN PARENT Albums ON DELETE CASCADE;
PostgreSQL
CREATE TABLE singers ( singer_id BIGINT PRIMARY KEY, first_name VARCHAR(1024), last_name VARCHAR(1024), singer_info BYTEA ); CREATE TABLE albums ( singer_id BIGINT, album_id BIGINT, album_title VARCHAR, PRIMARY KEY (singer_id, album_id) ) INTERLEAVE IN PARENT singers ON DELETE CASCADE; CREATE TABLE songs ( singer_id BIGINT, album_id BIGINT, track_id BIGINT, song_name VARCHAR, PRIMARY KEY (singer_id, album_id, track_id) ) INTERLEAVE IN PARENT albums ON DELETE CASCADE;
The following diagram represents a physical view of interleaved rows.
In this example, as the number of singers grows, Spanner adds split boundaries between singers to preserve data locality between a singer and its album and song data. However, if the size of a singer row and its child rows exceeds the split size limit, or a hotspot is detected in the child rows, Spanner attempts to add split boundaries to isolate that hotspot row along with all child rows below it.
In summary, a parent table along with all of its child and descendant tables forms a hierarchy of tables in the schema. Although each table in the hierarchy is logically independent, physically interleaving them this way can improve performance, effectively pre-joining the tables and allowing you to access related rows together while minimizing storage accesses.
Joins with interleaved tables
If possible, join data in interleaved tables by primary key. Because each
interleaved row is usually stored physically in the same split as its parent
row, Spanner can perform joins by primary key locally, minimizing
storage access and network traffic. In the following example,
Singers and Albums are joined on the primary key SingerId.
GoogleSQL
SELECT s.FirstName, a.AlbumTitle FROM Singers AS s JOIN Albums AS a ON s.SingerId = a.SingerId;
PostgreSQL
SELECT s.first_name, a.album_title FROM singers AS s JOIN albums AS a ON s.singer_id = a.singer_id;
Key columns
This section includes some notes about key columns.
Changing table keys
The keys of a table can't change; you can't add a key column to an existing table or remove a key column from an existing table.
Storing NULLs in a primary key
In GoogleSQL, if you
would like to store NULL in a primary key column, omit the NOT NULL clause
for that column in the schema. (PostgreSQL-dialect databases don't support NULLs in a primary
key column.)
Here's an example of omitting the NOT NULL clause on the primary key column
SingerId. Note that because SingerId is the primary key, there can be
only one row that stores NULL in that column.
CREATE TABLE Singers ( SingerId INT64, FirstName STRING(1024), LastName STRING(1024), ) PRIMARY KEY (SingerId);
The nullable property of the primary key column must match between the parent
and the child table declarations. In this example, NOT NULL for
the column Albums.SingerId is not allowed because Singers.SingerId omits it.
CREATE TABLE Singers ( SingerId INT64, FirstName STRING(1024), LastName STRING(1024), ) PRIMARY KEY (SingerId); CREATE TABLE Albums ( SingerId INT64 NOT NULL, AlbumId INT64 NOT NULL, AlbumTitle STRING(MAX), ) PRIMARY KEY (SingerId, AlbumId), INTERLEAVE IN PARENT Singers ON DELETE CASCADE;
Disallowed types
The following columns cannot be of type ARRAY:
- A table's key columns.
- An index's key columns.
Design for multi-tenancy
You might want to implement multi-tenancy if you are storing data that belongs to different customers. For example, a music service might want to store each individual record label's content separately.
Classic multi-tenancy
The classic way to design for multi-tenancy is to create a separate database
for each customer. In this example, each database has its own Singers table:
| SingerId | FirstName | LastName |
|---|---|---|
| 1 | Marc | Richards |
| 2 | Catalina | Smith |
| SingerId | FirstName | LastName |
|---|---|---|
| 1 | Alice | Trentor |
| 2 | Gabriel | Wright |
| SingerId | FirstName | LastName |
|---|---|---|
| 1 | Benjamin | Martinez |
| 2 | Hannah | Harris |
Schema-managed multi-tenancy
Another way to design for multi-tenancy in Spanner is to
have all customers in a single table in a single database, and to use a
different primary key value for each customer. For example, you could include a
CustomerId key column in your tables. If you make CustomerId the first key
column, then the data for each customer has good locality. Spanner
can then effectively use database splits to maximize
performance based on data size and load patterns. In the following example,
there is a single Singers table for all customers:
| CustomerId | SingerId | FirstName | LastName |
|---|---|---|---|
| 1 | 1 | Marc | Richards |
| 1 | 2 | Catalina | Smith |
| 2 | 1 | Alice | Trentor |
| 2 | 2 | Gabriel | Wright |
| 3 | 1 | Benjamin | Martinez |
| 3 | 2 | Hannah | Harris |
If you must have separate databases for each tenant, there are constraints to be aware of:
- There are limits on the number of databases per instance and the number of tables and indexes per database. Depending on the number of customers, it might not be possible to have separate databases or tables.
- Adding new tables and non-interleaved indexes can take a long time. You might not be able to get the performance you want if your schema design depends on adding new tables and indexes.
If you want to create separate databases, you might have more success if you distribute your tables across databases in such a way that each database has a low number of schema changes per week.
If you create separate tables and indexes for each customer of your application, do not put all of the tables and indexes in the same database. Instead, split them across many databases, to mitigate the performance issues with creating a large number of indexes.
To learn more about other data management patterns and application design for multi-tenancy, see Implementing Multi-Tenancy in Cloud Spanner