pg_ducklake

pig build pkg pg_ducklake         # 构建 RPM / DEB 包

pig repo add pgsql -u          # 添加仓库并更新缓存

shared_preload_libraries = 'pg_ducklake';

CREATE EXTENSION pg_ducklake;

CREATE EXTENSION pg_ducklake;

CREATE TABLE events (
  id int,
  kind text,
  ts timestamptz
) USING ducklake;

INSERT INTO events VALUES
  (1, 'login', now()),
  (2, 'click', now());

SELECT * FROM events ORDER BY id;

CREATE TABLE lake_events (
  id int,
  payload jsonb
) WITH (
  ducklake.table_path = 's3://my-bucket/prefix/'
) USING ducklake;

SELECT max(snapshot_id) AS before_delete
FROM ducklake.ducklake_snapshot \gset

DELETE FROM events WHERE id = 1;

SELECT * FROM ducklake.time_travel('events'::regclass, :before_delete);

CREATE TABLE row_store AS
SELECT i AS id, 'hello pg_ducklake' AS msg
FROM generate_series(1, 10000) AS i;

CREATE TABLE col_store USING ducklake AS
SELECT * FROM row_store;

CREATE TABLE titanic USING ducklake AS
SELECT * FROM ducklake.read_csv('https://raw.githubusercontent.com/datasciencedojo/datasets/master/titanic.csv');

CALL ducklake.set_option('data_inlining_row_limit', 100);
SELECT * FROM ducklake.flush_inlined_data('events'::regclass);

CALL ducklake.set_partition('events'::regclass, 'bucket(4, id)', 'month(ts)');
CREATE INDEX ON events USING ducklake_sorted (id, ts);

SELECT * FROM ducklake.merge_adjacent_files('events'::regclass);
CALL ducklake.set_option('expire_older_than', '7 days');
SELECT * FROM ducklake.expire_snapshots();
SELECT * FROM ducklake.cleanup_old_files();

INSTALL ducklake;
LOAD ducklake;
ATTACH 'ducklake:postgres:dbname=postgres host=localhost' AS my_ducklake
  (METADATA_SCHEMA 'ducklake');

SELECT * FROM my_ducklake.public.events;