A session in the Cloud Spanner API.

Output only. The approximate timestamp when the session is last used. It is typically earlier than the actual last use time.

Output only. The timestamp when the session is created.

The labels for the session.

• Label keys must be between 1 and 63 characters long and must conform to the following regular expression: [a-z]([-a-z0-9]*[a-z0-9])?.
• Label values must be between 0 and 63 characters long and must conform to the regular expression ([a-z]([-a-z0-9]*[a-z0-9])?)?.
• No more than 64 labels can be associated with a given session.

See https://goo.gl/xmQnxf for more information on and examples of labels.

The name of the session. This is always system-assigned; values provided when creating a session are ignored.

A session in the Cloud Spanner API.

Output only. The approximate timestamp when the session is last used. It is typically earlier than the actual last use time.

Output only. The timestamp when the session is created.

The labels for the session.

• Label keys must be between 1 and 63 characters long and must conform to the following regular expression: [a-z]([-a-z0-9]*[a-z0-9])?.
• Label values must be between 0 and 63 characters long and must conform to the regular expression ([a-z]([-a-z0-9]*[a-z0-9])?)?.
• No more than 64 labels can be associated with a given session.

See https://goo.gl/xmQnxf for more information on and examples of labels.

The name of the session. This is always system-assigned; values provided when creating a session are ignored.

An isolated set of Cloud Spanner resources on which databases can be hosted.

Required. The name of the instance's configuration. Values are of the form projects/<project>/instanceConfigs/<configuration>. See also InstanceConfig and ListInstanceConfigs.

Required. The descriptive name for this instance as it appears in UIs. Must be unique per project and between 4 and 30 characters in length.

Output only. A set of endpoint URIs based on your instance config that you can use instead of the global endpoint spanner.googleapis.com.

For example, if your instance config is us-central1 (a regional config in Iowa), then your instance specific endpoints may include us-central1-spanner.googleapis.com. By calling these endpoints instead of the global endpoint, you optimize network routing which could reduce network latency.

The client libraries, JDBC drivers, and other SDK clients automatically call these instance specific endpoints.

If you are using DNS whitelists, firewalls, or filtering to control access to endpoints, make sure you grant access to *spanner.googleapis.com.

Cloud Labels are a flexible and lightweight mechanism for organizing cloud resources into groups that reflect a customer's organizational needs and deployment strategies. Cloud Labels can be used to filter collections of resources. They can be used to control how resource metrics are aggregated. And they can be used as arguments to policy management rules (e.g. route, firewall, load balancing, etc.).

• Label keys must be between 1 and 63 characters long and must conform to the following regular expression: [a-z]([-a-z0-9]*[a-z0-9])?.
• Label values must be between 0 and 63 characters long and must conform to the regular expression ([a-z]([-a-z0-9]*[a-z0-9])?)?.
• No more than 64 labels can be associated with a given resource.

See https://goo.gl/xmQnxf for more information on and examples of labels.

If you plan to use labels in your own code, please note that additional characters may be allowed in the future. And so you are advised to use an internal label representation, such as JSON, which doesn't rely upon specific characters being disallowed. For example, representing labels as the string: name + "" + value would prove problematic if we were to allow "" in a future release.

Required. A unique identifier for the instance, which cannot be changed after the instance is created. Values are of the form projects/<project>/instances/a-z*[a-z0-9]. The final segment of the name must be between 2 and 64 characters in length.

The number of nodes allocated to this instance. This may be zero in API responses for instances that are not yet in state READY.

See the documentation for more information about nodes.

Output only. The current instance state. For CreateInstance, the state must be either omitted or set to CREATING. For UpdateInstance, the state must be either omitted or set to READY.

An isolated set of Cloud Spanner resources on which databases can be hosted.

Required. The name of the instance's configuration. Values are of the form projects/<project>/instanceConfigs/<configuration>. See also InstanceConfig and ListInstanceConfigs.

Required. The descriptive name for this instance as it appears in UIs. Must be unique per project and between 4 and 30 characters in length.

Output only. A set of endpoint URIs based on your instance config that you can use instead of the global endpoint spanner.googleapis.com.

For example, if your instance config is us-central1 (a regional config in Iowa), then your instance specific endpoints may include us-central1-spanner.googleapis.com. By calling these endpoints instead of the global endpoint, you optimize network routing which could reduce network latency.

The client libraries, JDBC drivers, and other SDK clients automatically call these instance specific endpoints.

If you are using DNS whitelists, firewalls, or filtering to control access to endpoints, make sure you grant access to *spanner.googleapis.com.

Cloud Labels are a flexible and lightweight mechanism for organizing cloud resources into groups that reflect a customer's organizational needs and deployment strategies. Cloud Labels can be used to filter collections of resources. They can be used to control how resource metrics are aggregated. And they can be used as arguments to policy management rules (e.g. route, firewall, load balancing, etc.).

• Label keys must be between 1 and 63 characters long and must conform to the following regular expression: [a-z]([-a-z0-9]*[a-z0-9])?.
• Label values must be between 0 and 63 characters long and must conform to the regular expression ([a-z]([-a-z0-9]*[a-z0-9])?)?.
• No more than 64 labels can be associated with a given resource.

See https://goo.gl/xmQnxf for more information on and examples of labels.

If you plan to use labels in your own code, please note that additional characters may be allowed in the future. And so you are advised to use an internal label representation, such as JSON, which doesn't rely upon specific characters being disallowed. For example, representing labels as the string: name + "" + value would prove problematic if we were to allow "" in a future release.

Required. A unique identifier for the instance, which cannot be changed after the instance is created. Values are of the form projects/<project>/instances/a-z*[a-z0-9]. The final segment of the name must be between 2 and 64 characters in length.

The number of nodes allocated to this instance. This may be zero in API responses for instances that are not yet in state READY.

See the documentation for more information about nodes.

Output only. The current instance state. For CreateInstance, the state must be either omitted or set to CREATING. For UpdateInstance, the state must be either omitted or set to READY.

Encapsulates settings provided to GetIamPolicy.

Optional. The policy format version to be returned.

Valid values are 0, 1, and 3. Requests specifying an invalid value will be rejected.

Requests for policies with any conditional bindings must specify version 3. Policies without any conditional bindings may specify any valid value or leave the field unset.

An Identity and Access Management (IAM) policy, which specifies access controls for Google Cloud resources.

A Policy is a collection of bindings. A binding binds one or more members to a single role. Members can be user accounts, service accounts, Google groups, and domains (such as G Suite). A role is a named list of permissions; each role can be an IAM predefined role or a user-created custom role.

Optionally, a binding can specify a condition, which is a logical expression that allows access to a resource only if the expression evaluates to true. A condition can add constraints based on attributes of the request, the resource, or both.

JSON example:

{
  "bindings": [
    {
      "role": "roles/resourcemanager.organizationAdmin",
      "members": [
        "user:mike@example.com",
        "group:admins@example.com",
        "domain:google.com",
        "serviceAccount:my-project-id@appspot.gserviceaccount.com"
      ]
    },
    {
      "role": "roles/resourcemanager.organizationViewer",
      "members": ["user:eve@example.com"],
      "condition": {
        "title": "expirable access",
        "description": "Does not grant access after Sep 2020",
        "expression": "request.time < timestamp('2020-10-01T00:00:00.000Z')",
      }
    }
  ],
  "etag": "BwWWja0YfJA=",
  "version": 3
}

YAML example:

bindings:
- members:
  - user:mike@example.com
  - group:admins@example.com
  - domain:google.com
  - serviceAccount:my-project-id@appspot.gserviceaccount.com
  role: roles/resourcemanager.organizationAdmin
- members:
  - user:eve@example.com
  role: roles/resourcemanager.organizationViewer
  condition:
    title: expirable access
    description: Does not grant access after Sep 2020
    expression: request.time < timestamp('2020-10-01T00:00:00.000Z')
- etag: BwWWja0YfJA=
- version: 3

For a description of IAM and its features, see the IAM documentation.

Associates a list of members to a role. Optionally, may specify a condition that determines how and when the bindings are applied. Each of the bindings must contain at least one member.

etag is used for optimistic concurrency control as a way to help prevent simultaneous updates of a policy from overwriting each other. It is strongly suggested that systems make use of the etag in the read-modify-write cycle to perform policy updates in order to avoid race conditions: An etag is returned in the response to getIamPolicy, and systems are expected to put that etag in the request to setIamPolicy to ensure that their change will be applied to the same version of the policy.

Important: If you use IAM Conditions, you must include the etag field whenever you call setIamPolicy. If you omit this field, then IAM allows you to overwrite a version 3 policy with a version 1 policy, and all of the conditions in the version 3 policy are lost.

Specifies the format of the policy.

Valid values are 0, 1, and 3. Requests that specify an invalid value are rejected.

Any operation that affects conditional role bindings must specify version 3. This requirement applies to the following operations:

• Getting a policy that includes a conditional role binding
• Adding a conditional role binding to a policy
• Changing a conditional role binding in a policy
• Removing any role binding, with or without a condition, from a policy that includes conditions

Important: If you use IAM Conditions, you must include the etag field whenever you call setIamPolicy. If you omit this field, then IAM allows you to overwrite a version 3 policy with a version 1 policy, and all of the conditions in the version 3 policy are lost.

If a policy does not include any conditions, operations on that policy may specify any valid version or leave the field unset.

Transactions

Each session can have at most one active transaction at a time. After the active transaction is completed, the session can immediately be re-used for the next transaction. It is not necessary to create a new session for each transaction.

Transaction Modes

Cloud Spanner supports three transaction modes:

1. Locking read-write. This type of transaction is the only way to write data into Cloud Spanner. These transactions rely on pessimistic locking and, if necessary, two-phase commit. Locking read-write transactions may abort, requiring the application to retry.

2. Snapshot read-only. This transaction type provides guaranteed consistency across several reads, but does not allow writes. Snapshot read-only transactions can be configured to read at timestamps in the past. Snapshot read-only transactions do not need to be committed.

3. Partitioned DML. This type of transaction is used to execute a single Partitioned DML statement. Partitioned DML partitions the key space and runs the DML statement over each partition in parallel using separate, internal transactions that commit independently. Partitioned DML transactions do not need to be committed.

For transactions that only read, snapshot read-only transactions provide simpler semantics and are almost always faster. In particular, read-only transactions do not take locks, so they do not conflict with read-write transactions. As a consequence of not taking locks, they also do not abort, so retry loops are not needed.

Transactions may only read/write data in a single database. They may, however, read/write data in different tables within that database.

Locking Read-Write Transactions

Locking transactions may be used to atomically read-modify-write data anywhere in a database. This type of transaction is externally consistent.

Clients should attempt to minimize the amount of time a transaction is active. Faster transactions commit with higher probability and cause less contention. Cloud Spanner attempts to keep read locks active as long as the transaction continues to do reads, and the transaction has not been terminated by Commit or Rollback. Long periods of inactivity at the client may cause Cloud Spanner to release a transaction's locks and abort it.

Conceptually, a read-write transaction consists of zero or more reads or SQL statements followed by Commit. At any time before Commit, the client can send a Rollback request to abort the transaction.

Semantics

Cloud Spanner can commit the transaction if all read locks it acquired are still valid at commit time, and it is able to acquire write locks for all writes. Cloud Spanner can abort the transaction for any reason. If a commit attempt returns ABORTED, Cloud Spanner guarantees that the transaction has not modified any user data in Cloud Spanner.

Unless the transaction commits, Cloud Spanner makes no guarantees about how long the transaction's locks were held for. It is an error to use Cloud Spanner locks for any sort of mutual exclusion other than between Cloud Spanner transactions themselves.

Retrying Aborted Transactions

When a transaction aborts, the application can choose to retry the whole transaction again. To maximize the chances of successfully committing the retry, the client should execute the retry in the same session as the original attempt. The original session's lock priority increases with each consecutive abort, meaning that each attempt has a slightly better chance of success than the previous.

Under some circumstances (e.g., many transactions attempting to modify the same row(s)), a transaction can abort many times in a short period before successfully committing. Thus, it is not a good idea to cap the number of retries a transaction can attempt; instead, it is better to limit the total amount of wall time spent retrying.

Idle Transactions

A transaction is considered idle if it has no outstanding reads or SQL queries and has not started a read or SQL query within the last 10 seconds. Idle transactions can be aborted by Cloud Spanner so that they don't hold on to locks indefinitely. In that case, the commit will fail with error ABORTED.

If this behavior is undesirable, periodically executing a simple SQL query in the transaction (e.g., SELECT 1) prevents the transaction from becoming idle.

Snapshot Read-Only Transactions

Snapshot read-only transactions provides a simpler method than locking read-write transactions for doing several consistent reads. However, this type of transaction does not support writes.

Snapshot transactions do not take locks. Instead, they work by choosing a Cloud Spanner timestamp, then executing all reads at that timestamp. Since they do not acquire locks, they do not block concurrent read-write transactions.

Unlike locking read-write transactions, snapshot read-only transactions never abort. They can fail if the chosen read timestamp is garbage collected; however, the default garbage collection policy is generous enough that most applications do not need to worry about this in practice.

Snapshot read-only transactions do not need to call Commit or Rollback (and in fact are not permitted to do so).

To execute a snapshot transaction, the client specifies a timestamp bound, which tells Cloud Spanner how to choose a read timestamp.

The types of timestamp bound are:

• Strong (the default).
• Bounded staleness.
• Exact staleness.

If the Cloud Spanner database to be read is geographically distributed, stale read-only transactions can execute more quickly than strong or read-write transaction, because they are able to execute far from the leader replica.

Each type of timestamp bound is discussed in detail below.

Strong

Strong reads are guaranteed to see the effects of all transactions that have committed before the start of the read. Furthermore, all rows yielded by a single read are consistent with each other -- if any part of the read observes a transaction, all parts of the read see the transaction.

Strong reads are not repeatable: two consecutive strong read-only transactions might return inconsistent results if there are concurrent writes. If consistency across reads is required, the reads should be executed within a transaction or at an exact read timestamp.

See TransactionOptions.ReadOnly.strong.

Exact Staleness

These timestamp bounds execute reads at a user-specified timestamp. Reads at a timestamp are guaranteed to see a consistent prefix of the global transaction history: they observe modifications done by all transactions with a commit timestamp <= the read timestamp, and observe none of the modifications done by transactions with a larger commit timestamp. They will block until all conflicting transactions that may be assigned commit timestamps <= the read timestamp have finished.

The timestamp can either be expressed as an absolute Cloud Spanner commit timestamp or a staleness relative to the current time.

These modes do not require a "negotiation phase" to pick a timestamp. As a result, they execute slightly faster than the equivalent boundedly stale concurrency modes. On the other hand, boundedly stale reads usually return fresher results.

See TransactionOptions.ReadOnly.read_timestamp and TransactionOptions.ReadOnly.exact_staleness.

Bounded Staleness

Bounded staleness modes allow Cloud Spanner to pick the read timestamp, subject to a user-provided staleness bound. Cloud Spanner chooses the newest timestamp within the staleness bound that allows execution of the reads at the closest available replica without blocking.

All rows yielded are consistent with each other -- if any part of the read observes a transaction, all parts of the read see the transaction. Boundedly stale reads are not repeatable: two stale reads, even if they use the same staleness bound, can execute at different timestamps and thus return inconsistent results.

Boundedly stale reads execute in two phases: the first phase negotiates a timestamp among all replicas needed to serve the read. In the second phase, reads are executed at the negotiated timestamp.

As a result of the two phase execution, bounded staleness reads are usually a little slower than comparable exact staleness reads. However, they are typically able to return fresher results, and are more likely to execute at the closest replica.

Because the timestamp negotiation requires up-front knowledge of which rows will be read, it can only be used with single-use read-only transactions.

See TransactionOptions.ReadOnly.max_staleness and TransactionOptions.ReadOnly.min_read_timestamp.

Old Read Timestamps and Garbage Collection

Cloud Spanner continuously garbage collects deleted and overwritten data in the background to reclaim storage space. This process is known as "version GC". By default, version GC reclaims versions after they are one hour old. Because of this, Cloud Spanner cannot perform reads at read timestamps more than one hour in the past. This restriction also applies to in-progress reads and/or SQL queries whose timestamp become too old while executing. Reads and SQL queries with too-old read timestamps fail with the error FAILED_PRECONDITION.

Partitioned DML Transactions

Partitioned DML transactions are used to execute DML statements with a different execution strategy that provides different, and often better, scalability properties for large, table-wide operations than DML in a ReadWrite transaction. Smaller scoped statements, such as an OLTP workload, should prefer using ReadWrite transactions.

Partitioned DML partitions the keyspace and runs the DML statement on each partition in separate, internal transactions. These transactions commit automatically when complete, and run independently from one another.

To reduce lock contention, this execution strategy only acquires read locks on rows that match the WHERE clause of the statement. Additionally, the smaller per-partition transactions hold locks for less time.

That said, Partitioned DML is not a drop-in replacement for standard DML used in ReadWrite transactions.

• The DML statement must be fully-partitionable. Specifically, the statement must be expressible as the union of many statements which each access only a single row of the table.

• The statement is not applied atomically to all rows of the table. Rather, the statement is applied atomically to partitions of the table, in independent transactions. Secondary index rows are updated atomically with the base table rows.

• Partitioned DML does not guarantee exactly-once execution semantics against a partition. The statement will be applied at least once to each partition. It is strongly recommended that the DML statement should be idempotent to avoid unexpected results. For instance, it is potentially dangerous to run a statement such as UPDATE table SET column = column + 1 as it could be run multiple times against some rows.

• The partitions are committed automatically - there is no support for Commit or Rollback. If the call returns an error, or if the client issuing the ExecuteSql call dies, it is possible that some rows had the statement executed on them successfully. It is also possible that statement was never executed against other rows.

• Partitioned DML transactions may only contain the execution of a single DML statement via ExecuteSql or ExecuteStreamingSql.

• If any error is encountered during the execution of the partitioned DML operation (for instance, a UNIQUE INDEX violation, division by zero, or a value that cannot be stored due to schema constraints), then the operation is stopped at that point and an error is returned. It is possible that at this point, some partitions have been committed (or even committed multiple times), and other partitions have not been run at all.

Given the above, Partitioned DML is good fit for large, database-wide, operations that are idempotent, such as deleting old rows from a very large table.

Partitioned DML transaction.

Authorization to begin a Partitioned DML transaction requires spanner.databases.beginPartitionedDmlTransaction permission on the session resource.

Transaction will not write.

Authorization to begin a read-only transaction requires spanner.databases.beginReadOnlyTransaction permission on the session resource.

Transaction may write.

Authorization to begin a read-write transaction requires spanner.databases.beginOrRollbackReadWriteTransaction permission on the session resource.

A modification to one or more Cloud Spanner rows. Mutations can be applied to a Cloud Spanner database by sending them in a Commit call.

Delete rows from a table. Succeeds whether or not the named rows were present.

Insert new rows in a table. If any of the rows already exist, the write or transaction fails with error ALREADY_EXISTS.

Like insert, except that if the row already exists, then its column values are overwritten with the ones provided. Any column values not explicitly written are preserved.

When using insert_or_update, just as when using insert, all NOT NULL columns in the table must be given a value. This holds true even when the row already exists and will therefore actually be updated.

Like insert, except that if the row already exists, it is deleted, and the column values provided are inserted instead. Unlike insert_or_update, this means any values not explicitly written become NULL.

In an interleaved table, if you create the child table with the ON DELETE CASCADE annotation, then replacing a parent row also deletes the child rows. Otherwise, you must delete the child rows before you replace the parent row.

Update existing rows in a table. If any of the rows does not already exist, the transaction fails with error NOT_FOUND.

Transactions

Each session can have at most one active transaction at a time. After the active transaction is completed, the session can immediately be re-used for the next transaction. It is not necessary to create a new session for each transaction.

Transaction Modes

Cloud Spanner supports three transaction modes:

1. Locking read-write. This type of transaction is the only way to write data into Cloud Spanner. These transactions rely on pessimistic locking and, if necessary, two-phase commit. Locking read-write transactions may abort, requiring the application to retry.

2. Snapshot read-only. This transaction type provides guaranteed consistency across several reads, but does not allow writes. Snapshot read-only transactions can be configured to read at timestamps in the past. Snapshot read-only transactions do not need to be committed.

3. Partitioned DML. This type of transaction is used to execute a single Partitioned DML statement. Partitioned DML partitions the key space and runs the DML statement over each partition in parallel using separate, internal transactions that commit independently. Partitioned DML transactions do not need to be committed.

For transactions that only read, snapshot read-only transactions provide simpler semantics and are almost always faster. In particular, read-only transactions do not take locks, so they do not conflict with read-write transactions. As a consequence of not taking locks, they also do not abort, so retry loops are not needed.

Transactions may only read/write data in a single database. They may, however, read/write data in different tables within that database.

Locking Read-Write Transactions

Locking transactions may be used to atomically read-modify-write data anywhere in a database. This type of transaction is externally consistent.

Clients should attempt to minimize the amount of time a transaction is active. Faster transactions commit with higher probability and cause less contention. Cloud Spanner attempts to keep read locks active as long as the transaction continues to do reads, and the transaction has not been terminated by Commit or Rollback. Long periods of inactivity at the client may cause Cloud Spanner to release a transaction's locks and abort it.

Conceptually, a read-write transaction consists of zero or more reads or SQL statements followed by Commit. At any time before Commit, the client can send a Rollback request to abort the transaction.

Semantics

Cloud Spanner can commit the transaction if all read locks it acquired are still valid at commit time, and it is able to acquire write locks for all writes. Cloud Spanner can abort the transaction for any reason. If a commit attempt returns ABORTED, Cloud Spanner guarantees that the transaction has not modified any user data in Cloud Spanner.

Unless the transaction commits, Cloud Spanner makes no guarantees about how long the transaction's locks were held for. It is an error to use Cloud Spanner locks for any sort of mutual exclusion other than between Cloud Spanner transactions themselves.

Retrying Aborted Transactions

When a transaction aborts, the application can choose to retry the whole transaction again. To maximize the chances of successfully committing the retry, the client should execute the retry in the same session as the original attempt. The original session's lock priority increases with each consecutive abort, meaning that each attempt has a slightly better chance of success than the previous.

Under some circumstances (e.g., many transactions attempting to modify the same row(s)), a transaction can abort many times in a short period before successfully committing. Thus, it is not a good idea to cap the number of retries a transaction can attempt; instead, it is better to limit the total amount of wall time spent retrying.

Idle Transactions

A transaction is considered idle if it has no outstanding reads or SQL queries and has not started a read or SQL query within the last 10 seconds. Idle transactions can be aborted by Cloud Spanner so that they don't hold on to locks indefinitely. In that case, the commit will fail with error ABORTED.

If this behavior is undesirable, periodically executing a simple SQL query in the transaction (e.g., SELECT 1) prevents the transaction from becoming idle.

Snapshot Read-Only Transactions

Snapshot read-only transactions provides a simpler method than locking read-write transactions for doing several consistent reads. However, this type of transaction does not support writes.

Snapshot transactions do not take locks. Instead, they work by choosing a Cloud Spanner timestamp, then executing all reads at that timestamp. Since they do not acquire locks, they do not block concurrent read-write transactions.

Unlike locking read-write transactions, snapshot read-only transactions never abort. They can fail if the chosen read timestamp is garbage collected; however, the default garbage collection policy is generous enough that most applications do not need to worry about this in practice.

Snapshot read-only transactions do not need to call Commit or Rollback (and in fact are not permitted to do so).

To execute a snapshot transaction, the client specifies a timestamp bound, which tells Cloud Spanner how to choose a read timestamp.

The types of timestamp bound are:

• Strong (the default).
• Bounded staleness.
• Exact staleness.

If the Cloud Spanner database to be read is geographically distributed, stale read-only transactions can execute more quickly than strong or read-write transaction, because they are able to execute far from the leader replica.

Each type of timestamp bound is discussed in detail below.

Strong

Strong reads are guaranteed to see the effects of all transactions that have committed before the start of the read. Furthermore, all rows yielded by a single read are consistent with each other -- if any part of the read observes a transaction, all parts of the read see the transaction.

Strong reads are not repeatable: two consecutive strong read-only transactions might return inconsistent results if there are concurrent writes. If consistency across reads is required, the reads should be executed within a transaction or at an exact read timestamp.

See TransactionOptions.ReadOnly.strong.

Exact Staleness

These timestamp bounds execute reads at a user-specified timestamp. Reads at a timestamp are guaranteed to see a consistent prefix of the global transaction history: they observe modifications done by all transactions with a commit timestamp <= the read timestamp, and observe none of the modifications done by transactions with a larger commit timestamp. They will block until all conflicting transactions that may be assigned commit timestamps <= the read timestamp have finished.

The timestamp can either be expressed as an absolute Cloud Spanner commit timestamp or a staleness relative to the current time.

These modes do not require a "negotiation phase" to pick a timestamp. As a result, they execute slightly faster than the equivalent boundedly stale concurrency modes. On the other hand, boundedly stale reads usually return fresher results.

See TransactionOptions.ReadOnly.read_timestamp and TransactionOptions.ReadOnly.exact_staleness.

Bounded Staleness

Bounded staleness modes allow Cloud Spanner to pick the read timestamp, subject to a user-provided staleness bound. Cloud Spanner chooses the newest timestamp within the staleness bound that allows execution of the reads at the closest available replica without blocking.

All rows yielded are consistent with each other -- if any part of the read observes a transaction, all parts of the read see the transaction. Boundedly stale reads are not repeatable: two stale reads, even if they use the same staleness bound, can execute at different timestamps and thus return inconsistent results.

Boundedly stale reads execute in two phases: the first phase negotiates a timestamp among all replicas needed to serve the read. In the second phase, reads are executed at the negotiated timestamp.

As a result of the two phase execution, bounded staleness reads are usually a little slower than comparable exact staleness reads. However, they are typically able to return fresher results, and are more likely to execute at the closest replica.

Because the timestamp negotiation requires up-front knowledge of which rows will be read, it can only be used with single-use read-only transactions.

See TransactionOptions.ReadOnly.max_staleness and TransactionOptions.ReadOnly.min_read_timestamp.

Old Read Timestamps and Garbage Collection

Cloud Spanner continuously garbage collects deleted and overwritten data in the background to reclaim storage space. This process is known as "version GC". By default, version GC reclaims versions after they are one hour old. Because of this, Cloud Spanner cannot perform reads at read timestamps more than one hour in the past. This restriction also applies to in-progress reads and/or SQL queries whose timestamp become too old while executing. Reads and SQL queries with too-old read timestamps fail with the error FAILED_PRECONDITION.

Partitioned DML Transactions

Partitioned DML transactions are used to execute DML statements with a different execution strategy that provides different, and often better, scalability properties for large, table-wide operations than DML in a ReadWrite transaction. Smaller scoped statements, such as an OLTP workload, should prefer using ReadWrite transactions.

Partitioned DML partitions the keyspace and runs the DML statement on each partition in separate, internal transactions. These transactions commit automatically when complete, and run independently from one another.

To reduce lock contention, this execution strategy only acquires read locks on rows that match the WHERE clause of the statement. Additionally, the smaller per-partition transactions hold locks for less time.

That said, Partitioned DML is not a drop-in replacement for standard DML used in ReadWrite transactions.

• The DML statement must be fully-partitionable. Specifically, the statement must be expressible as the union of many statements which each access only a single row of the table.

• The statement is not applied atomically to all rows of the table. Rather, the statement is applied atomically to partitions of the table, in independent transactions. Secondary index rows are updated atomically with the base table rows.

• Partitioned DML does not guarantee exactly-once execution semantics against a partition. The statement will be applied at least once to each partition. It is strongly recommended that the DML statement should be idempotent to avoid unexpected results. For instance, it is potentially dangerous to run a statement such as UPDATE table SET column = column + 1 as it could be run multiple times against some rows.

• The partitions are committed automatically - there is no support for Commit or Rollback. If the call returns an error, or if the client issuing the ExecuteSql call dies, it is possible that some rows had the statement executed on them successfully. It is also possible that statement was never executed against other rows.

• Partitioned DML transactions may only contain the execution of a single DML statement via ExecuteSql or ExecuteStreamingSql.

• If any error is encountered during the execution of the partitioned DML operation (for instance, a UNIQUE INDEX violation, division by zero, or a value that cannot be stored due to schema constraints), then the operation is stopped at that point and an error is returned. It is possible that at this point, some partitions have been committed (or even committed multiple times), and other partitions have not been run at all.

Given the above, Partitioned DML is good fit for large, database-wide, operations that are idempotent, such as deleting old rows from a very large table.

Partitioned DML transaction.

Authorization to begin a Partitioned DML transaction requires spanner.databases.beginPartitionedDmlTransaction permission on the session resource.

Transaction will not write.

Authorization to begin a read-only transaction requires spanner.databases.beginReadOnlyTransaction permission on the session resource.

Transaction may write.

Authorization to begin a read-write transaction requires spanner.databases.beginOrRollbackReadWriteTransaction permission on the session resource.

A single DML statement.

It is not always possible for Cloud Spanner to infer the right SQL type from a JSON value. For example, values of type BYTES and values of type STRING both appear in params as JSON strings.

In these cases, param_types can be used to specify the exact SQL type for some or all of the SQL statement parameters. See the definition of Type for more information about SQL types.

Parameter names and values that bind to placeholders in the DML string.

A parameter placeholder consists of the @ character followed by the parameter name (for example, @firstName). Parameter names can contain letters, numbers, and underscores.

Parameters can appear anywhere that a literal value is expected. The same parameter name can be used more than once, for example:

"WHERE id > @msg_id AND id < @msg_id + 100"

It is an error to execute a SQL statement with unbound parameters.

Required. The DML string.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

Options for a PartitionQueryRequest and PartitionReadRequest.

Note: This hint is currently ignored by PartitionQuery and PartitionRead requests.

The desired maximum number of partitions to return. For example, this may be set to the number of workers available. The default for this option is currently 10,000. The maximum value is currently 200,000. This is only a hint. The actual number of partitions returned may be smaller or larger than this maximum count request.

Note: This hint is currently ignored by PartitionQuery and PartitionRead requests.

The desired data size for each partition generated. The default for this option is currently 1 GiB. This is only a hint. The actual size of each partition may be smaller or larger than this size request.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

KeySet defines a collection of Cloud Spanner keys and/or key ranges. All the keys are expected to be in the same table or index. The keys need not be sorted in any particular way.

If the same key is specified multiple times in the set (for example if two ranges, two keys, or a key and a range overlap), Cloud Spanner behaves as if the key were only specified once.

For convenience all can be set to true to indicate that this KeySet matches all keys in the table or index. Note that any keys specified in keys or ranges are only yielded once.

A list of specific keys. Entries in keys should have exactly as many elements as there are columns in the primary or index key with which this KeySet is used. Individual key values are encoded as described here.

A list of key ranges. See KeyRange for more information about key range specifications.

Options for a PartitionQueryRequest and PartitionReadRequest.

Note: This hint is currently ignored by PartitionQuery and PartitionRead requests.

The desired maximum number of partitions to return. For example, this may be set to the number of workers available. The default for this option is currently 10,000. The maximum value is currently 200,000. This is only a hint. The actual number of partitions returned may be smaller or larger than this maximum count request.

Note: This hint is currently ignored by PartitionQuery and PartitionRead requests.

The desired data size for each partition generated. The default for this option is currently 1 GiB. This is only a hint. The actual size of each partition may be smaller or larger than this size request.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

KeySet defines a collection of Cloud Spanner keys and/or key ranges. All the keys are expected to be in the same table or index. The keys need not be sorted in any particular way.

If the same key is specified multiple times in the set (for example if two ranges, two keys, or a key and a range overlap), Cloud Spanner behaves as if the key were only specified once.

For convenience all can be set to true to indicate that this KeySet matches all keys in the table or index. Note that any keys specified in keys or ranges are only yielded once.

A list of specific keys. Entries in keys should have exactly as many elements as there are columns in the primary or index key with which this KeySet is used. Individual key values are encoded as described here.

A list of key ranges. See KeyRange for more information about key range specifications.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.

KeySet defines a collection of Cloud Spanner keys and/or key ranges. All the keys are expected to be in the same table or index. The keys need not be sorted in any particular way.

If the same key is specified multiple times in the set (for example if two ranges, two keys, or a key and a range overlap), Cloud Spanner behaves as if the key were only specified once.

For convenience all can be set to true to indicate that this KeySet matches all keys in the table or index. Note that any keys specified in keys or ranges are only yielded once.

A list of specific keys. Entries in keys should have exactly as many elements as there are columns in the primary or index key with which this KeySet is used. Individual key values are encoded as described here.

A list of key ranges. See KeyRange for more information about key range specifications.

This message is used to select the transaction in which a Read or ExecuteSql call runs.

See TransactionOptions for more information about transactions.

Begin a new transaction and execute this read or SQL query in it. The transaction ID of the new transaction is returned in ResultSetMetadata.transaction, which is a Transaction.

Execute the read or SQL query in a previously-started transaction.

Execute the read or SQL query in a temporary transaction. This is the most efficient way to execute a transaction that consists of a single SQL query.