BigQuery から ClickHouse へのデータの読み込み

DECLARE export_path STRING;
DECLARE n INT64;
DECLARE i INT64;
SET i = 0;

-- We recommend setting n to correspond to x billion rows. So 5 billion rows, n = 5
SET n = 100;

WHILE i < n DO
  SET export_path = CONCAT('gs://mybucket/mytable/', i,'-*.parquet');
  EXPORT DATA
    OPTIONS (
      uri = export_path,
      format = 'PARQUET',
      overwrite = true
    )
  AS (
    SELECT * FROM mytable WHERE export_id = i
  );
  SET i = i + 1;
END WHILE;

-- If your BigQuery table contains a column of type STRUCT, you must enable this setting
-- to map that column to a ClickHouse column of type Nested
SET input_format_parquet_import_nested = 1;

CREATE TABLE default.mytable
(
        `timestamp` DateTime64(6),
        `some_text` String
)
ENGINE = MergeTree
ORDER BY (timestamp);

SET parallel_distributed_insert_select = 1;

INSERT INTO mytable
SELECT
    timestamp,
    ifNull(some_text, '') AS some_text
FROM s3Cluster(
    'default',
    'https://storage.googleapis.com/mybucket/mytable/*.parquet.gz',
    '<ACCESS_ID>',
    '<SECRET>'
);

SELECT * FROM mytable LIMIT 10;