INSERT INTO

INSERT INTO

sql
    INSERT INTO iceberg_catalog.db.mirror_table
    SELECT * FROM internal_catalog.db.upsert_table;

INSERT

UPDATE

DELETE

INSERT INTO

default_catalog

iceberg_glue

sql
CREATE EXTERNAL CATALOG iceberg_glue
PROPERTIES (
    "type" = "iceberg",
    "iceberg.catalog.type" = "glue",
    "aws.glue.region" = "us-east-1",
    "aws.s3.endpoint" = "<http://s3.us-east-1.amazonaws.com|s3.us-east-1.amazonaws.com>"
);

catalog.database.mv_name

sql
CREATE MATERIALIZED VIEW iceberg_glue.target_db.mirror_table
PARTITION BY (dt) -- Partitioning helps with incremental mirroring
DISTRIBUTED BY HASH(id)
REFRESH ASYNC EVERY (INTERVAL 10 MINUTE) -- Set your sync frequency
PROPERTIES (
    "storage_format" = "parquet"
)
AS 
SELECT 
    id, 
    user_name, 
    event_time, 
    date_trunc('day', event_time) as dt
FROM default_catalog.native_db.fast_upsert_table;

REFRESH

sql
    REFRESH MATERIALIZED VIEW iceberg_glue.target_db.mirror_table;

INSERT INTO

INSERT INTO iceberg_glue.target_db.table SELECT ...

starrocks/cn-ubuntu:4.1.3

VARIANT

sql
CREATE EXTERNAL CATALOG iceberg_glue
PROPERTIES (
    "type" = "iceberg",
    "iceberg.catalog.type" = "glue",
    "aws.glue.region" = "us-east-1"
);

sql
CREATE MATERIALIZED VIEW native_db.app_mirror_table
-- Optimize for your query patterns
DISTRIBUTED BY HASH(user_id) 
-- Partition by the same column as Iceberg for incremental refreshes
PARTITION BY (event_date) 
REFRESH ASYNC EVERY (INTERVAL 5 MINUTE) 
PROPERTIES (
    "replication_num" = "3",
    "storage_medium" = "SSD"
)
AS 
SELECT 
    user_id, 
    event_time, 
    payload,
    CAST(event_time AS DATE) as event_date
FROM iceberg_glue.source_db.raw_data;

native_db.app_mirror_table

"storage_medium" = "SSD"

"replication_num"

sql
    REFRESH MATERIALIZED VIEW native_db.app_mirror_table;

sql
    SELECT * FROM information_schema.task_runs
    WHERE task_name = 'mv_app_mirror_table'
    ORDER BY create_time DESC;

CREATE MATERIALIZED VIEW iceberg.db.test_mirror
DISTRIBUTED BY HASH(id)
REFRESH ASYNC EVERY (INTERVAL 1 MINUTE) -- Set your sync frequency
PROPERTIES (
    "storage_format" = "parquet"
)
AS
SELECT
	*
FROM default_catalog.starrocks.test_mirror;

set catalog <default_catalog>

default_catalog

iceberg

INSERT INTO

INSERT INTO

TASK

sql
-- Switch to your iceberg catalog
SET CATALOG iceberg_glue;

CREATE TABLE target_db.mirrored_table (
    id BIGINT,
    data STRING,
    dt DATE
) PARTITION BY (dt);

TASK

SELECT

INSERT

sql
-- Tasks must be created in the default_catalog
CREATE TASK task_mirror_to_iceberg
SCHEDULE EVERY 1 MINUTE -- Set your sync frequency here
AS
INSERT INTO iceberg_glue.target_db.mirrored_table
SELECT 
    id, 
    data, 
    dt
FROM default_catalog.native_db.native_table
-- Optional: Add logic here to only sync new data (Incremental sync)
-- WHERE update_time > (SELECT max(update_time) FROM iceberg_glue.target_db.mirrored_table)
;

DEFERRED

sql
ALTER TASK task_mirror_to_iceberg RESUME;

INSERT INTO

WHERE

sql
SELECT * FROM information_schema.task_runs 
WHERE task_name = 'task_mirror_to_iceberg' 
ORDER BY create_time DESC;

1785201840.698829

*"How could i stand up a starrocks shared nothing database that mirrors data i have in an external iceberg catalog..."*

*"i want to have a table in my iceberg catalog be mirrored from a starrocks native table..."*

default_catalog

iceberg

INSERT INTO

sql
-- This creates the automation you are looking for
CREATE TASK mirror_to_iceberg
SCHEDULE EVERY 5 MINUTE  -- Your desired sync interval
AS
INSERT INTO iceberg.db.test_mirror
SELECT * FROM default_catalog.starrocks.test_mirror;

-- Start the background process
ALTER TASK mirror_to_iceberg RESUME;

INSERT INTO