Information schema for PostgreSQL-dialect databases

The information schema is a built-in schema that's common to every PostgreSQL database. You can run SQL queries against tables in the information_schema to fetch schema metadata for a database.

For example, the following query fetches the names of all user-defined tables in a database:

  SELECT
    table_name
  FROM
    information_schema.tables
  WHERE
    table_schema = 'public'

Usage

  • information_schema tables are available only through SQL interfaces, for example:

    • The executeQuery API
    • The gcloud spanner databases execute-sql command
    • The Query page of a database in the Google Cloud console.

    Other single read methods do not support information_schema.

Differences from information_schema for PostgreSQL

The tables in the information_schema for PostgreSQL-dialect databases include columns from the tables in the information_schema for open-source PostgreSQL and in some cases also include columns from Spanner. In these tables, the open-source PostgreSQL columns come first and in the same order as they do for a open-source PostgreSQL database, and any distinct columns for Spanner are appended afterwards. Queries written for the open-source PostgreSQL version of information_schema should work without modification when using PostgreSQL-dialect databases in Google Cloud CLI.

Other notable differences in the information_schema for PostgreSQL-dialect databases are:

  • Some of the table columns for open-source PostgreSQL are available, but not populated in PostgreSQL-dialect databases.
  • PostgreSQL-dialect databases use public for the default schema name.
  • Automatically generated constraint names use a different format than open-source PostgreSQL databases.
  • Tables related to open-source PostgreSQL features that are not supported in PostgreSQL-dialect databases are not available.
  • Some tables that are available with Spanner but not open-source PostgreSQL, such as database_options, index_columns, indexes, and spanner_statistics are available.

Tables in information_schema for PostgreSQL-dialect databases

The tables and views in the information_schema are compatible with the tables and views in the information_schema of open-source PostgreSQL.

The following sections describe the tables and views in the information_schema for PostgreSQL-dialect databases:

change_stream_columns

This view contains the relationship between a database's columns and its change streams. Each row regards one change stream and one database column.

The data in change_stream_columns does not include the implicit relationships between columns and change streams that track those columns' entire tables.

Column name Type Description
change_stream_catalog character varying The database name.
change_stream_schema character varying The name of the change stream's schema. The default is `public` for PostgreSQL-dialect databases.
change_stream_name character varying The name of the change stream.
table_catalog character varying The database name.
table_schema character varying The name of the table's schema. The default is `public` for PostgreSQL-dialect databases.
table_name character varying The name of the table that this row refers to.
column_name character varying The name of the column that this row refers to.

change_stream_options

This view contains the configuration options for change streams.

Column name Type Description
change_stream_catalog character varying The database name.
change_stream_schema character varying The name of the change stream's schema. The default is `public` for PostgreSQL-dialect databases.
change_stream_name character varying The name of the change stream.
option_name character varying The name of the change stream option.
option_type character varying The data type of the change stream option.
option_value character varying The value of the change stream option.

change_stream_tables

This view contains the relationships between a database's tables and its change streams. Each row regards one database table and one change stream.

The data in change_stream_tables does not include the implicit relationships between tables and change streams that track the entire database.

Column name Type Description
change_stream_catalog character varying The database name.
change_stream_schema character varying The name of the change stream's schema. The default is `public` for PostgreSQL-dialect databases.
change_stream_name character varying The name of the change stream that this row refers to.
table_catalog character varying The database name.
table_schema character varying The name of the table's schema. The default is `public` for PostgreSQL-dialect databases.
table_name character varying The name of the table that this row refers to.
all_columns character varying YES if this row's change stream tracks the entirety of the table this row refers to. Otherwise, NO. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.

change_streams

This view lists all of a database's change streams, and notes which ones track the entire database versus specific tables or columns.

Column name Type Description
change_stream_catalog character varying The database name.
change_stream_schema character varying The name of this change stream's schema. The default is `public` for PostgreSQL-dialect databases.
change_stream_name character varying The name of the change stream.
all character varying YES if this change stream tracks the entire database. NO if this change stream tracks specific tables or columns. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.

check_constraints

The check_constraints view contains one row for each check constraint defined by either the CHECK or the NOT NULL keyword.

Column name Type Description
constraint_catalog character varying The database name.
constraint_schema character varying The name of the constraint's schema. The default is `public` for PostgreSQL-dialect databases.
constraint_name character varying The name of the constraint. If the name of the constraint is not explicitly specified in the schema, the auto-generated name is used.
check_clause character varying The check constraint's expression.
spanner_state character varying The current state of the check constraint. The possible states are as follows:
  • VALIDATING: The PostgreSQL-dialect database is validating the existing data for an `ALTER CONSTRAINT` or `ADD CONSTRAINT` command.
  • COMMITTED: There is no active schema change for this constraint.

column_column_usage

This view lists all the generated columns that depend on another base column in the same table.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the table. The name is `public` for the default schema and non-empty for other schemas (for example, the information_schema itself). This column is never null.
table_name character varying The name of the table that contains the generated columns.
column_name character varying The name of the base column that the generated column depends on.
dependent_column character varying The name of the generated column.

column_options

This view lists all the options defined for the referenced table columns of a foreign key constraint. The view contains only those columns in the reference table that the current user has access to (by way of being the owner or granted privileges).

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the foreign table. The name is `public` for the default schema and non-empty for other schemas (for example, the information_schema itself). This column is never null.
table_name character varying The name of the foreign table.
column_name character varying The name of the column.
option_name character varying A SQL identifier that uniquely identifies the option. This identifier is the key of the OPTIONS clause in DDL.
option_value character varying A SQL literal describing the value of this option. The value of this column is parsable as part of a query.
option_type character varying A data type name that is the type of this option value.

columns

This view provides information about all table columns (or view columns) in the database. The view contains only those columns that the current user has access to (by way of being the owner or granted privileges).

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the table. The name is `public` for the default schema and non-empty for other schemas (for example, the information_schema itself). This column is never null.
table_name character varying The name of the table
column_name character varying The name of the column
ordinal_position BIGINT The ordinal position of the column in the table, starting with a value of 1
column_default character varying A string representation of the open-source PostgreSQL expression of the default value of the column, for example, '9'::bigint.
is_nullable character varying A string that indicates whether the column is nullable. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.
data_type character varying The data type of the column. The value is one of the following:
  • For built-in types, the name of the data type.
  • For arrays, the value ARRAY.
character_maximum_length BIGINT The declared maximum length for character and bit string data types. If a maximum length was not specified, then the value is `NULL`. If the data type of the column is not a character or bit string, then the value is `NULL`.
character_octet_length BIGINT Not currently used. The value is always `NULL`.
numeric_precision BIGINT The precision of the numeric data type of the current column. For `double precision`, the value is 53. For `bigint`, the value is 64. For all other data types, the value is `NULL`.
numeric_precision_radix BIGINT The base (unit) of the precision for numeric types. Currently only two values are supported:
  • 2 for `double precision`, `float8`, and `bigint`
  • 10 for `numeric`
For all other data types the value is `NULL`.
numeric_scale BIGINT Contains the scale of the numeric column type, which is the number of precision base units after the radix point. For `bigint`, the value is 0. For all other data types, the value is `NULL`.
datetime_precision BIGINT Not currently used. The value is always `NULL`.
interval_type character varying Not currently used. The value is always `NULL`.
interval_precision BIGINT Not currently used. The value is always `NULL`.
character_set_catalog character varying Not currently used. The value is always `NULL`.
character_set_schema character varying Not currently used. The value is always `NULL`.
character_set_name character varying Not currently used. The value is always `NULL`.
collation_catalog character varying Not currently used. The value is always `NULL`.
collation_schema character varying Not currently used. The value is always `NULL`.
collation_name character varying Not currently used. The value is always `NULL`.
domain_catalog character varying Not currently used. The value is always `NULL`.
domain_schema character varying Not currently used. The value is always `NULL`.
domain_name character varying Not currently used. The value is always `NULL`.
udt_catalog character varying Not currently used. The value is always `NULL`.
udt_schema character varying Not currently used. The value is always `NULL`.
udt_name character varying Not currently used. The value is always `NULL`.
scope_catalog character varying Not currently used. The value is always `NULL`.
scope_schema character varying Not currently used. The value is always `NULL`.
scope_name character varying Not currently used. The value is always `NULL`.
maximum_cardinality BIGINT Not currently used. The value is always `NULL`.
dtd_identifier character varying Not currently used. The value is always `NULL`.
is_self_referencing character varying Not currently used. The value is always `NULL`.
is_identity character varying Not currently used. The value is always `NULL`.
identity_generation character varying Not currently used. The value is always `NULL`.
identity_start character varying Not currently used. The value is always `NULL`.
identity_increment character varying Not currently used. The value is always `NULL`.
identity_maximum character varying Not currently used. The value is always `NULL`.
identity_minimum character varying Not currently used. The value is always `NULL`.
identity_cycle character varying Not currently used. The value is always `NULL`.
is_generated character varying A string that indicates whether the column is generated. The string is either ALWAYS for a generated column or NEVER for a non-generated column.
generation_expression character varying A string representing the SQL expression of a generated column, or NULL if the column is not a generated column.
is_updatable character varying Not currently used. The value is always `NULL`.
spanner_type character varying A string holding the DDL-compatible type of the column.
is_stored character varying A string that indicates whether the generated column is stored. The string is always YES or NO for generated columns, and NULL for non-generated columns.
spanner_state character varying The current state of the column. A new stored generated column added to an existing table may go through multiple user-observable states before it is fully usable. Possible values are:
  • NO_WRITE: No read or write is allowed to the columns. A stored generated column in this state does not cause any client effect.
  • WRITE_ONLY: The column is being backfilled. No read is allowed.
  • COMMITTED: The column is fully usable.
  • NULL: Used for columns in system schemas.

constraint_column_usage

This view contains one row about each column used by a constraint.

  • For PRIMARY KEY and CHECK constraints defined by the NOT NULL keyword, the view contains those columns.
  • For CHECK constraints created with the CHECK keyword, the view includes the columns used by the check constraint expression.
  • For foreign key constraints, the view contains the columns of the referenced table.
  • For UNIQUE constraints, the view contains the columns from KEY_COLUMN_USAGE.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the table that contains the column that is used by the constraint.
table_name character varying The name of the table that contains the column that is used by the constraint.
column_name character varying The name of the column that is used by the constraint.
constraint_catalog character varying The database name.
constraint_schema character varying The name of the constraint's schema.
constraint_name character varying The name of the constraint.

constraint_table_usage

This view contains one row for each table used by a constraint. For FOREIGN KEY constraints, the table information is for the tables in the REFERENCES clause. For a unique or primary key constraint, this view simply identifies the table the constraint belongs to. Check constraints and not-null constraints are not included in this view.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the constrained table's schema.
table_name character varying The name of the table that is used by some constraint.
constraint_catalog character varying The database name.
constraint_schema character varying The name of the schema that contains the constraint.
constraint_name character varying The name of the constraint.

database_options

This table lists the options that are set on the database.

Column name Type Description
catalog_name character varying The database name.
schema_name character varying The name of the schema. The default value is `public` for PostgreSQL-dialect databases.
option_name character varying The name of the database option. This is the value of `key` in the `OPTIONS` clause in DDL.
option_type character varying The data type of the database option.
option_value character varying The value of the database option.

index_columns

This table lists the columns in an index.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the index. The default value is `public`.
table_name character varying The name of the table associated with the index.
index_name character varying The name of the index. Tables with a `PRIMARY KEY` specification have a pseudo-index entry generated with the name `PRIMARY_KEY`.
index_type character varying The type of index. Possible values are `PRIMARY_KEY`, `LOCAL`, or `GLOBAL`.
column_name character varying The name of the column.
ordinal_position BIGINT The ordinal position of the column in the index (or primary key), starting with a value of 1. This value is NULL for non-key columns (for example, columns specified in the INCLUDE clause of an index).
column_ordering character varying The sort order of the column. The value is ASC or DESC for key columns, and NULL for non-key columns (for example, columns specified in the STORING clause of an index).
is_nullable character varying A string that indicates whether the column is nullable. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.
spanner_type character varying A string holding the DDL-compatible type of the column.

indexes

This view lists the indexes in a schema.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema. The default value is `public`.
table_name character varying The name of the table.
index_name character varying The name of the index. Tables created with a PRIMARY KEY clause have a pseudo-index entry generated with the name PRIMARY_KEY, which allows the fields of the primary key to be identified.
index_type character varying The type of the index. The values include PRIMARY_KEY, LOCAL, or GLOBAL.
parent_table_name character varying Secondary indexes can be interleaved in a parent table, as discussed in Creating a secondary index. This column holds the name of that parent table, or NULL if the index is not interleaved.
is_unique character varying Whether the index keys must be unique. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.
is_null_filtered character varying Whether the index includes entries with NULL values. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.
index_state character varying The current state of the index. Possible values and the states they represent are:
  • NULL: the index type is `PRIMARY_KEY`
  • PREPARE: creating empty tables for a new index
  • WRITE_ONLY: backfilling data for a new index
  • WRITE_ONLY_CLEANUP: cleaning up a new index
  • WRITE_ONLY_VALIDATE_UNIQUE: checking uniqueness of data in a new index
  • READ_WRITE: normal index operation
spanner_is_managed character varying Whether the index is managed by Spanner. For example, secondary backing indexes for foreign keys are managed by Spanner. The string is either YES or NO, rather than a Boolean value, in accordance with the SQL standard.

information_schema_catalog_name

This table contains one row and one column containing the database name.

Column name Type Description
catalog_name character varying The database name.

key_column_usage

This view identifies all columns in the current database that are referenced by a unique, primary key, or foreign key constraint. For information about CHECK constraint columns, see the check_constraints view.

Column name Type Description
constraint_catalog character varying The database name.
constraint_schema character varying The name of the constraint's schema. The default value is `public`.
constraint_name character varying The name of the constraint.
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the table that contains the constrained column. The default value is `public`.
table_name character varying The name of the table that contains the column that is restricted by this constraint.
column_name character varying The name of the column that is constrained.
ordinal_position BIGINT The ordinal position of the column within the constraint's key, starting with a value of 1.
position_in_unique_constraint BIGINT For FOREIGN KEYs, the ordinal position of the column within the unique constraint, starting with a value of 1. This column has a `NULL` value for other constraint types.

referential_constraints

This view contains one row about each FOREIGN KEY constraint. You can see only those constraints for which you have write access to the referencing table. This view also identifies the PRIMARY KEY and UNIQUE constraints on the referenced tables that the foreign keys use for constraint enforcement and referential actions.

Column name Type Description
constraint_catalog character varying The database name.
constraint_schema character varying The name of the schema that contains the foreign key constraint. The default value is `public`.
constraint_name character varying The name of the foreign key constraint.
unique_constraint_catalog character varying The database name.
unique_constraint_schema character varying The name of the schema that contains the unique or primary key constraint that the foreign key constraint references.
unique_constraint_name character varying The name of the unique or primary key constraint that the foreign key constraint references.
match_option character varying The match method used by the foreign key constraint. The value is always NONE.
update_rule character varying The update rule of the foreign key constraint. This value is always NO ACTION.
delete_rule character varying The delete rule of the foreign key constraint. This value is always NO ACTION.
spanner_state character varying The current state of the foreign key. Spanner does not begin enforcing the constraint until the foreign key's backing indexes are created and backfilled. Once the indexes are ready, Spanner begins enforcing the constraint for new transactions while it validates the existing data. The possible values and the states they represent are:
  • BACKFILLING_INDEXES: Indexes are being backfilled.
  • VALIDATING_DATA: Existing data and new writes are being validated.
  • WAITING_FOR_COMMIT: The foreign key bulk operations have completed successfully, or none were needed, but the foreign key is still pending.
  • COMMITTED: The schema change was committed.

schemata

The information_schema.schemata view contains one row for each schema in the current database. The schemas include the information schema and a default schema named public.

Column name Type Description
catalog_name character varying The database name.
schema_name character varying The name of the schema. This is set to `public` for the default schema and non-empty for named schemas.
schema_owner character varying The name of the owner of the schema.
default_character_set_catalog character varying Not currently used.
default_character_set_schema character varying Not currently used.
default_character_set_name character varying Not currently used.
sql_path character varying Not currently used.
effective_timestamp timestamp with timezone The timestamp at which all the data in this schema became effective. This is currently used only for the default schema.

spanner_statistics

This table lists the available query optimizer statistics packages.

Column name Type Description
catalog_name character varying The database name.
schema_name character varying The name of the schema. The default schema value is `public`.
package_name character varying The name of the statistics package.
allow_gc character varying Whether the statistics package is exempted from garbage collection. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value. This attribute must be set to NO before you can reference the statistics package in a hint or through the client API.

table_constraints

This view contains all constraints belonging to tables that the current user has access to (other than SELECT).

Column name Type Description
constraint_catalog character varying The database name.
constraint_schema character varying The name of the schema that contains the constraint.
constraint_name character varying The name of the constraint.
table_catalog character varying The database name.
table_schema character varying The name of schema that contains the table associated with the constraint.
table_name character varying The name of the table.
constraint_type character varying The type of the constraint. Possible values are:
  • CHECK
  • FOREIGN KEY
  • PRIMARY KEY
  • UNIQUE
is_deferrable character varying The value is always NO.
initially_deferred character varying The value is always NO.
enforced character varying Whether the constraint is enforced. If a constraint is enforced, (after it reaches a certain state), it is validated both at write time and by a background integrity verifier. In accordance with the SQL standard, the string is either YES or NO, rather than a Boolean value.

tables

This view contains all the tables and views in the current database.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema that contains the table or view.
table_name character varying The name of the table or view.
table_type character varying The table type. Possible values include 'BASE TABLE' and 'VIEW'.
self_referencing_column_name character varying Not currently used.
reference_generation character varying Not currently used.
user_defined_type_catalog character varying Not currently used.
user_defined_type_schema character varying Not currently used.
user_defined_type_name character varying Not currently used.
is_insertable_into character varying Not currently used.
is_typed character varying Not currently used.
commit_action character varying Not currently used.
parent_table_name character varying The name of the parent table if this table is interleaved, or NULL.
on_delete_action character varying This is set to CASCADE or NO ACTION for interleaved tables, and NULL otherwise. See TABLE statements for more information.
spanner_state character varying The current creation state of the table.
A table can go through multiple states during creation, if bulk operations are involved, for example, when the table is created with a foreign key that requires backfilling of its referenced index. Possible states are:
  • ADDING_FOREIGN_KEY: Adding the table's foreign keys
  • WAITING_FOR_COMMIT: Finalizing the schema change
  • COMMITTED: The schema change to create the table has been committed. You cannot write to the table until the change is committed.
  • NULL: Tables or views that are not base tables.
interleave_type character varying Whether there exists a parent-child relationship between this table and the table it is interleaved in. Possible values are:
  • IN: An `INTERLEAVE IN` table with no parent-child relationship. A row in this table can exist regardless of the existence of its parent table row.
  • IN PARENT: An `INTERLEAVE IN PARENT` table with a parent-child relationship. A row in this table requires the existence of its parent table row.
row_deletion_policy_expression character varying An string that contains the expression text that defines the ROW DELETION POLICY.

views

This view lists information about the views in a schema.

Column name Type Description
table_catalog character varying The database name.
table_schema character varying The name of the schema. The default value is `public`.
table_name character varying The name of the view.
view_definition character varying The SQL text of the query that defines the view.
check_option character varying Not currently used.
is_updatable character varying Not currently used.
is_insertable_into character varying Not currently used.
is_trigger_updatable character varying Not currently used.
is_trigger_deletable character varying Not currently used.
is_trigger_insertable_into character varying Not currently used.

Examples

Return information about each table in the default schema:

SELECT
  t.table_catalog,
  t.table_name,
  t.parent_table_name
FROM
  information_schema.tables AS t
WHERE
  t.table_schema = 'public'
ORDER BY
  t.table_catalog,
  t.table_schema,
  t.table_name

Return the name of all tables and views in the information_schema for PostgreSQL-dialect databases:

SELECT table_name
FROM information_schema.tables
WHERE table_schema = "information_schema"

Return information about the columns in the user table my_table:

SELECT
  t.ordinal_position,
  t.column_name,
  t.data_type,
  t.spanner_type,
  t.is_nullable
FROM
  information_schema.columns AS t
WHERE
  t.table_schema = 'public'
  AND
  t.table_name = 'my_table'
ORDER BY
  t.ordinal_position

Return information about each index in the default schema in the current database:

SELECT
  t.table_name,
  t.index_name,
  t.parent_table_name
FROM
  information_schema.indexes AS t
WHERE
  t.table_schema = 'public'
  AND
  t.index_type != 'PRIMARY_KEY'
ORDER BY
  t.table_schema,
  t.table_name,
  t.index_name

Return all the columns that use options other than the default:

SELECT
  t.table_name,
  t.column_name,
  t.option_type,
  t.option_value,
  t.option_name
FROM
  information_schema.column_options AS t
WHERE
  t.table_schema = 'public'
ORDER BY
  t.table_schema,
  t.table_name,
  t.column_name,
  t.option_name

Return the current optimizer-related database options: 

SELECT
  s.option_name,
  s.option_value
FROM
  information_schema.database_options s
WHERE
  s.schema_name='public'
  AND s.option_name IN ('optimizer_version',
    'optimizer_statistics_package')

Return all available statistics packages: 

SELECT *
FROM information_schema.spanner_statistics;

What's next