CREATE STREAM¶

Creates a new stream in the current/specified schema or replaces an existing stream. A stream records data manipulation language (DML) changes made to a table, directory table, external table, or the underlying tables in a view (including secure views). The object for which changes are recorded is called the source object.

In addition, this command supports the following variant:

  • CREATE STREAM … CLONE (creates a clone of an existing stream)

See also:

ALTER STREAM , DROP STREAM , SHOW STREAMS , DESCRIBE STREAM

Syntax¶

The command syntax differs depending on the object on which the stream is created:

-- table
CREATE [ OR REPLACE ] STREAM [IF NOT EXISTS]
  <name>
  [ COPY GRANTS ]
  ON TABLE <table_name>
  [ { AT | BEFORE } ( { TIMESTAMP => <timestamp> | OFFSET => <time_difference> | STATEMENT => <id> | STREAM => '<name>' } ) ]
  [ APPEND_ONLY = TRUE | FALSE ]
  [ SHOW_INITIAL_ROWS = TRUE | FALSE ]
  [ COMMENT = '<string_literal>' ]

-- External table
CREATE [ OR REPLACE ] STREAM [IF NOT EXISTS]
  <name>
  [ COPY GRANTS ]
  ON EXTERNAL TABLE <external_table_name>
  [ { AT | BEFORE } ( { TIMESTAMP => <timestamp> | OFFSET => <time_difference> | STATEMENT => <id> | STREAM => '<name>' } ) ]
  [ INSERT_ONLY = TRUE ]
  [ COMMENT = '<string_literal>' ]

-- Directory table
CREATE [ OR REPLACE ] STREAM [IF NOT EXISTS]
  <name>
  [ COPY GRANTS ]
  ON STAGE <stage_name>
  [ COMMENT = '<string_literal>' ]

-- View
CREATE [ OR REPLACE ] STREAM [IF NOT EXISTS]
  <name>
  [ COPY GRANTS ]
  ON VIEW <view_name>
  [ { AT | BEFORE } ( { TIMESTAMP => <timestamp> | OFFSET => <time_difference> | STATEMENT => <id> | STREAM => '<name>' } ) ]
  [ APPEND_ONLY = TRUE | FALSE ]
  [ SHOW_INITIAL_ROWS = TRUE | FALSE ]
  [ COMMENT = '<string_literal>' ]
Copy

Variant Syntax¶

CREATE STREAM … CLONE

Creates a new stream with the same definition as the source stream. The clone inherits the current offset (i.e. the current transactional table version) from the source stream.

CREATE [ OR REPLACE ] STREAM <name> CLONE <source_stream>
  [ COPY GRANTS ]
  [ ... ]
Copy

For more information about cloning, see CREATE <object> … CLONE.

Required Parameters¶

name

String that specifies the identifier (i.e. name) for the stream; must be unique for the schema in which the stream is created.

In addition, the identifier must start with an alphabetic character and cannot contain spaces or special characters unless the entire identifier string is enclosed in double quotes (e.g. "My object"). Identifiers enclosed in double quotes are also case-sensitive.

For more details, see Identifier requirements.

table_name

String that specifies the identifier (i.e. name) for the table whose changes are tracked by the stream (i.e. the source table).

Access control:

To query a stream, a role must have the SELECT privilege on the underlying table.

external_table_name

String that specifies the identifier (i.e. name) for the external table whose changes are tracked by the stream (i.e. the source external table).

Access control:

To query a stream, a role must have the SELECT privilege on the underlying external table.

stage_name

String that specifies the identifier (i.e. name) for the stage whose directory table changes are tracked by the stream (i.e. the source directory table).

Access control:

To query a stream, a role must have the USAGE (external stage) or READ (internal stage) privilege on the underlying stage.

view_name

String that specifies the identifier (i.e. name) for the source view. The stream tracks DML changes to the underlying tables in the view.

For more information about streams on views, see Streams on Views.

Access control:

To query a stream, a role must have the SELECT privilege on the view.

Optional Parameters¶

COPY GRANTS

Specifies to retain the access permissions from the original stream when a new stream is created using any of the following CREATE STREAM variants:

  • CREATE OR REPLACE STREAM

  • CREATE STREAM … CLONE

The parameter copies all permissions, except OWNERSHIP, from the existing stream to the new stream. By default, the role that executes the CREATE STREAM command owns the new stream.

Note

  • If the CREATE STREAM statement references more than one stream (e.g. create or replace stream t1 clone t2;), the COPY GRANTS clause gives precedence to the stream being replaced.

  • The SHOW GRANTS output for the replacement stream lists the grantee for the copied privileges as the role that executed the CREATE STREAM statement, with the current timestamp when the statement was executed.

  • The operation to copy grants occurs atomically in the CREATE STREAM command (i.e. within the same transaction).

Note

This parameter is not supported currently.

AT ( { TIMESTAMP => <timestamp> | OFFSET => <time_difference> | STATEMENT => <id> | STREAM => '<name>' ) | BEFORE ( TIMESTAMP => <timestamp> | OFFSET => <time_difference> | STATEMENT => <id> ) }

Creates a stream at a specific time/point in the past (using Time Travel). The AT | BEFORE clause determines the point in the past from which historical data is requested:

  • The AT keyword specifies that the request is inclusive of any changes made by a statement or transaction with a timestamp equal to the specified parameter.

    The STREAM => '<name>' value is special. When provided, the CREATE STREAM statement creates the new stream at the same offset as the specified stream. You can also provide this value when recreating an existing stream (using the OR REPLACE keywords) to retain the current offset of the stream after it is recreated. '<name>' is the identifier (i.e. name) for the existing stream whose offset is copied to the new or recreated stream.

    The new or recreated stream advances the offset, as usual, when the stream is used in a DML transaction.

  • The BEFORE keyword specifies that the request refers to a point immediately preceding the specified parameter.

Note

If no change tracking data is available on the source object at the point in the past specified in the AT | BEFORE clause, the CREATE STREAM statement fails. No stream can be created at a time in the past before change tracking was recorded.

APPEND_ONLY = TRUE | FALSE

Only supported for streams on standard tables or streams on views that query standard tables.

Specifies whether this is an append-only stream. Append-only streams track row inserts only. Update and delete operations (including table truncates) are not recorded. For example, if 10 rows are inserted into a table and then 5 of those rows are deleted before the offset for an append-only stream is advanced, the stream records 10 rows.

This type of stream improves query performance over standard streams and is very useful for extract, load, transform (ELT) and similar scenarios that depend exclusively on row inserts.

A standard stream joins the deleted and inserted rows in the change set to determine which rows were deleted and which were updated. An append-only stream returns the appended rows only and therefore can be much more performant than a standard stream. For example, the source table can be truncated immediately after the rows in an append-only stream are consumed, and the record deletions do not contribute to the overhead the next time the stream is queried or consumed.

Default:

FALSE

INSERT_ONLY = TRUE | FALSE

Required for streams on external tables. Not supported by streams on other objects.

Specifies whether this is an insert-only stream. Insert-only streams track row inserts only; they do not record delete operations that remove rows from an inserted set (i.e. no-ops). For example, in-between any two offsets, if File1 is removed from the cloud storage location referenced by the external table, and File2 is added, the stream returns records for the rows in File2 only. Unlike when tracking change data capture (CDC) data for standard tables, Snowflake cannot access the historical records for files in cloud storage.

Default:

FALSE

SHOW_INITIAL_ROWS = TRUE | FALSE

Specifies the records to return the first time the stream is consumed.

TRUE

The stream returns only the rows that existed in the source object at the moment when the stream was created. The METADATA$ISUPDATE column shows a FALSE value in these rows. Subsequently, the stream returns any DML changes to the source object since the most recent offset; that is, the normal stream behavior.

This parameter enables initializing any downstream process with the contents of the source object for the stream.

FALSE

The stream returns any DML changes to the source object since the most recent offset.

Default:

FALSE

COMMENT = 'string_literal'

String (literal) that specifies a comment for the stream.

Default: No value

Output¶

The output for a stream includes the same columns as the source object along with the following additional columns:

  • METADATA$ACTION: Specifies the action (INSERT or DELETE).

  • METADATA$ISUPDATE: Specifies whether the action recorded (INSERT or DELETE) is part of an UPDATE applied to the rows in the source table or view.

    Note that streams record the differences between two offsets. If a row is added and then updated in the current offset, the delta change is a new row. The METADATA$ISUPDATE row records a FALSE value.

  • METADATA$ROW_ID: Specifies the unique and immutable ID for the row, which can be used to track changes to specific rows over time.

Access Control Requirements¶

A role used to execute this SQL command must have the following privileges at a minimum:

Streams on standard tables:

Object

Privilege

Notes

Schema

CREATE STREAM

Table

SELECT

If change tracking has not been enabled on the source table (using ALTER TABLE … SET CHANGE_TRACKING = TRUE), then only the table owner (i.e. the role that has the OWNERSHIP privilege on the table) can create the initial stream on the table. Creating the initial stream automatically enables change tracking on the table.

Note that operating on any object in a schema also requires the USAGE privilege on the parent database and schema.

Streams on views:

Object

Privilege

Notes

Schema

CREATE STREAM

View

SELECT

If change tracking has not been enabled on the source view and its underlying tables, then only a role that has the OWNERSHIP privilege on the view and its underlying tables owner can create the initial stream on the view. Creating the initial stream automatically enables change tracking on the table. For instructions on enabling change tracking on a view and its underlying tables, refer to Enabling Change Tracking on Views and Underlying Tables. Note that enabling change tracking locks the underlying tables while change tracking is being enabled. Locks on the underlying objects may cause latency in DDL/DML operations with these objects. For more information, refer to Resource Locking.

Note that operating on any object in a schema also requires the USAGE privilege on the parent database and schema.

Streams on directory tables:

Object

Privilege

Notes

Schema

CREATE STREAM

Stage

USAGE (external stage) or READ (internal stage)

Note that operating on any object in a schema also requires the USAGE privilege on the parent database and schema.

Streams on external tables:

Object

Privilege

Notes

Schema

CREATE STREAM

External table

SELECT

Note that operating on any object in a schema also requires the USAGE privilege on the parent database and schema.

For instructions on creating a custom role with a specified set of privileges, see Creating custom roles.

For general information about roles and privilege grants for performing SQL actions on securable objects, see Overview of Access Control.

Usage Notes¶

  • A stream can be queried multiple times to update multiple objects in the same transaction and it will return the same data.

  • The stream position (i.e. offset) is advanced when the stream is used in a DML statement. The position is updated at the end of the transaction to the beginning timestamp of the transaction. The stream describes change records starting from the current position of the stream and ending at the current transactional timestamp.

    To ensure multiple statements access the same change records in the stream, surround them with an explicit transaction statement (BEGIN .. COMMIT). An explicit transaction locks the stream, so that DML updates to the source object are not reported to the stream until the transaction is committed.

  • Streams have no Fail-safe period or Time Travel retention period. The metadata in these objects cannot be recovered if a stream is dropped.

  • Streams on shared tables:

    • The retention period for a source table is not extended automatically to prevent any streams on the table from becoming stale.

  • Standard streams cannot retrieve change data for geospatial data. We recommend creating append-only streams on objects that contain geospatial data.

  • Streams on views:

    • Creating the first stream on a view using the view owner role (i.e. the role with the OWNERSHIP privilege on the view) enables change tracking on the view. If the same role also owns the underlying tables, change tracking is also enabled on the tables. If the role was not granted the OWNERSHIP privilege on both the view and its underlying tables, then change tracking must be enabled manually on the applicable objects. For instructions, see Enabling Change Tracking on Views and Underlying Tables.

    • Depending on the number of joins in a view, a single change in the underlying tables could result in a large number of changes in the stream output.

    • Any stream on a given view breaks if the source view or underlying tables are dropped or recreated (using CREATE OR REPLACE VIEW).

    • Any streams on a secure view adhere to the secure view constraints.

      If the owner of a non-secure view (i.e. the role with the OWNERSHIP privilege on the view) changes it to a secure view (using ALTER VIEW … SET SECURE), any stream on the view automatically enforces secure view constraints.

      In addition, the retention period for the underlying tables is not extended automatically to prevent any streams on the secure view from becoming stale.

    • Streams based on views where the view uses non-deterministic functions can return non-deterministic results.

      For example, the results of context functions such as CURRENT_DATE, and CURRENT_USER are non-deterministic. The results of data generation functions such as RANDOM are also non-deterministic. If a view contains a non-deterministic function, then any stream on that view will not be a constant snapshot of the function’s output. Instead the value in the stream may change when queried.

      We recommend that you ensure that the non-determinism in the results of a view does not affect the correctness of the stream query results.

      For an example, see Stream on a View That Calls a Non-deterministic SQL Function.

  • Streams on directory tables: The METADATA$ROW_ID column values in the stream output are empty.

  • Regarding metadata:

    Attention

    Customers should ensure that no personal data (other than for a User object), sensitive data, export-controlled data, or other regulated data is entered as metadata when using the Snowflake service. For more information, see Metadata Fields in Snowflake.

  • CREATE OR REPLACE <object> statements are atomic. That is, when an object is replaced, the old object is deleted and the new object is created in a single transaction.

Examples¶

Creating a Table Stream¶

Create a stream on the mytable table:

CREATE STREAM mystream ON TABLE mytable;
Copy

Using Time Travel with the Source Table¶

Create a stream on the mytable table as it existed before the date and time in the specified timestamp:

CREATE STREAM mystream ON TABLE mytable BEFORE (TIMESTAMP => TO_TIMESTAMP(40*365*86400));
Copy

Create a stream on the mytable table as it existed exactly at the date and time of the specified timestamp:

CREATE STREAM mystream ON TABLE mytable AT (TIMESTAMP => TO_TIMESTAMP_TZ('02/02/2019 01:02:03', 'mm/dd/yyyy hh24:mi:ss'));
Copy

Create a stream on the mytable table as it existed 5 minutes ago:

CREATE STREAM mystream ON TABLE mytable AT(OFFSET => -60*5);
Copy

Create a stream on the mytable table with the same offset as existing stream oldstream on the same source table:

CREATE STREAM mystream ON TABLE mytable AT(STREAM => 'oldstream');
Copy

Recreate the existing mystream stream but retain its current offset:

CREATE OR REPLACE STREAM mystream ON TABLE mytable AT(STREAM => 'mystream');
Copy

Create a stream on the mytable table including transactions up to, but not including any changes made by the specified transaction:

CREATE STREAM mystream ON TABLE mytable BEFORE(STATEMENT => '8e5d0ca9-005e-44e6-b858-a8f5b37c5726');
Copy

Creating a Stream on a Single-table View¶

Create a stream on the myview view:

CREATE STREAM mystream ON VIEW myview;
Copy

For additional examples, see Stream Examples.

Creating an Insert-only Stream on an External Table¶

Create an external table stream and query the change data capture records in the stream, which track the records added to the external table metadata:

-- Create an external table that points to the MY_EXT_STAGE stage.
-- The external table is partitioned by the date (in YYYY/MM/DD format) in the file path.
CREATE EXTERNAL TABLE my_ext_table (
  date_part date as to_date(substr(metadata$filename, 1, 10), 'YYYY/MM/DD'),
  ts timestamp AS (value:time::timestamp),
  user_id varchar AS (value:userId::varchar),
  color varchar AS (value:color::varchar)
) PARTITION BY (date_part)
  LOCATION=@my_ext_stage
  AUTO_REFRESH = false
  FILE_FORMAT=(TYPE=JSON);

-- Create a stream on the external table
CREATE STREAM my_ext_table_stream ON EXTERNAL TABLE my_ext_table INSERT_ONLY = TRUE;

-- Execute SHOW streams
-- The MODE column indicates that the new stream is an INSERT_ONLY stream
SHOW STREAMS;
+-------------------------------+------------------------+---------------+-------------+--------------+-----------+------------------------------------+-------+-------+-------------+
| created_on                    | name                   | database_name | schema_name | owner        | comment   | table_name                         | type  | stale | mode        |
|-------------------------------+------------------------+---------------+-------------+--------------+-----------+------------------------------------+-------+-------+-------------|
| 2020-08-02 05:13:20.174 -0800 | MY_EXT_TABLE_STREAM    | MYDB          | PUBLIC      | MYROLE       |           | MYDB.PUBLIC.EXTTABLE_S3_PART       | DELTA | false | INSERT_ONLY |
+-------------------------------+------------------------+---------------+-------------+--------------+-----------+------------------------------------+-------+-------+-------------+

-- Add a file named '2020/08/05/1408/log-08051409.json' to the stage using the appropriate tool for the cloud storage service.

-- Manually refresh the external table metadata.
ALTER EXTERNAL TABLE my_ext_table REFRESH;

-- Query the external table stream.
-- The stream indicates that the rows in the added JSON file were recorded in the external table metadata.
SELECT * FROM my_ext_table_stream;
+----------------------------------------+------------+-------------------------+---------+-------+-----------------+-------------------+-----------------+---------------------------------------------+
| VALUE                                  | DATE_PART  | TS                      | USER_ID | COLOR | METADATA$ACTION | METADATA$ISUPDATE | METADATA$ROW_ID | METADATA$FILENAME                           |
|----------------------------------------+------------+-------------------------+---------+-------+-----------------+-------------------+-----------------+---------------------------------------------|
| {                                      | 2020-08-05 | 2020-08-05 15:57:01.000 | user25  | green | INSERT          | False             |                 | test/logs/2020/08/05/1408/log-08051409.json |
|   "color": "green",                    |            |                         |         |       |                 |                   |                 |                                             |
|   "time": "2020-08-05 15:57:01-07:00", |            |                         |         |       |                 |                   |                 |                                             |
|   "userId": "user25"                   |            |                         |         |       |                 |                   |                 |                                             |
| }                                      |            |                         |         |       |                 |                   |                 |                                             |
| {                                      | 2020-08-05 | 2020-08-05 15:58:02.000 | user56  | brown | INSERT          | False             |                 | test/logs/2020/08/05/1408/log-08051409.json |
|   "color": "brown",                    |            |                         |         |       |                 |                   |                 |                                             |
|   "time": "2020-08-05 15:58:02-07:00", |            |                         |         |       |                 |                   |                 |                                             |
|   "userId": "user56"                   |            |                         |         |       |                 |                   |                 |                                             |
| }                                      |            |                         |         |       |                 |                   |                 |                                             |
+----------------------------------------+------------+-------------------------+---------+-------+-----------------+-------------------+-----------------+---------------------------------------------+
Copy

Creating a Standard Stream on a Directory Table¶

Create a stream on the directory table for a stage named mystage:

CREATE STREAM dirtable_mystage_s ON STAGE mystage;
Copy

Manually refresh the directory table metadata to populate the stream:

ALTER STAGE mystage REFRESH;
Copy

Query the stream after one or more files were added to the stage after the most recent offset for the stream:

SELECT * FROM dirtable_mystage_s;

+-------------------+--------+-------------------------------+----------------------------------+----------------------------------+-------------------------------------------------------------------------------------------+-----------------+-------------------+-----------------+
| RELATIVE_PATH     | SIZE   | LAST_MODIFIED                 | MD5                              | ETAG                             | FILE_URL                                                                                  | METADATA$ACTION | METADATA$ISUPDATE | METADATA$ROW_ID |
|-------------------+--------+-------------------------------+----------------------------------+----------------------------------+-------------------------------------------------------------------------------------------+-----------------+-------------------+-----------------|
| file1.csv.gz      |   1048 | 2021-05-14 06:09:08.000 -0700 | c98f600c492c39bef249e2fcc7a4b6fe | c98f600c492c39bef249e2fcc7a4b6fe | https://myaccount.snowflakecomputing.com/api/files/MYDB/MYSCHEMA/MYSTAGE/file1%2ecsv%2egz | INSERT          | False             |                 |
| file2.csv.gz      |   3495 | 2021-05-14 06:09:09.000 -0700 | 7f1a4f98ef4c7c42a2974504d11b0e20 | 7f1a4f98ef4c7c42a2974504d11b0e20 | https://myaccount.snowflakecomputing.com/api/files/MYDB/MYSCHEMA/MYSTAGE/file2%2ecsv%2egz | INSERT          | False             |                 |
+-------------------+--------+-------------------------------+----------------------------------+----------------------------------+-------------------------------------------------------------------------------------------+-----------------+-------------------+-----------------+
Copy