Presto

trino(presto) nested json 를 처리하기 위한 table schema 예제 trino(presto) - create table row : json object (nested json) 타입을 위한 sub column 정의 array : 배열 Example {"id":1,"name":"Alice"} {"id":2,"name":"Bob"} {"id":3,"name":"Carol"} {"id":4,"name":"David", "etc1": [{"key": "key-data11"}, {"key": "key-data22"}] } {"id":5,"name":"Elise", "etc2": ["data1", "data2", "data3"]} create table hive.default.sample ( id varchar, name varchar, etc1 array(row(key varchar)), etc2 array(varchar) ) with ( format = 'json', external_location = 's3a://data/sample/' ) trino cli trino> use hive.default; USE trino:default> select * from sample; id | name | etc1 | etc2 ----+-------+--------------------------------------+----------------------- 1 | Alice | NULL | NULL 2 | Bob | NULL | NULL 3 | Carol | NULL | NULL 4 | David | [{key=key-data11}, {key=key-data22}] | NULL 5 | Elise | NULL | [data1, data2, data3] (5 rows) Query 20220824_000627_00008_qpvtj, FINISHED, 1 node Splits: 1 total, 1 done (100.00%) 0.27 [5 rows, 212B] [18 rows/s, 777B/s]

trino(presto), minio docker 활용 테스트 환경 trino (hive connector) to minio trino 의 hive connector 를 활용해서 minio 데이터 분석 환경 Hive connector : https://trino.io/docs/current/connector/hive.html{:target="_blank"} 참고 : https://github.com/cdecl/trino-minio-docker{:target="_blank"} docker-compose (trino, minio) version: '3' services: trino: image: trinodb/trino container_name: trino restart: always ports: - "8080:8080" volumes: - ./etc:/etc/trino minio: image: minio/minio restart: always command: server /data --console-address ":9001" container_name: minio environment: MINIO_ROOT_USER: minio MINIO_ROOT_PASSWORD: minio1234 restart: always ports: - "9000:9000" - "9001:9001" volumes: - ./minio-data:/data minio (hive) properties etc/catalog/hive.properties metastore 는 minio catalog 버킷에 저장 s3a://catalog/trino/ minio 세팅 정보 : access-key, secret-key, endpoint connector.name=hive-hadoop2 hive.metastore=file hive.metastore.catalog.dir=s3a://catalog/trino/ hive.recursive-directories=true hive.non-managed-table-writes-enabled=true hive.allow-drop-table=true hive.s3.path-style-access=true hive.s3.ssl.enabled=false hive.s3select-pushdown.enabled=true hive.s3.aws-access-key=minio hive.s3.aws-secret-key=minio1234 hive.s3.endpoint=http://minio:9000 docker-compose 실행 및 초기값 세팅 ## 실행 $ docker-compose up -d reating network "trino_default" with the default driver Creating minio ... done Creating trino ... done rclone 으로 minio 버킷 생성 # rclone 설정 $ cat local.conf [local] type = s3 provider = Minio env_auth = false access_key_id = minio secret_access_key = minio1234 endpoint = http://localhost:9000 # create catalog, data directory $ rclone --config local.conf mkdir local:catalog $ rclone --config local.conf mkdir local:data # sample json (ndjson) data) $ cat 1.json {"id":1,"name":"Alice"} {"id":2,"name":"Bob"} {"id":3,"name":"Carol"} # test data copy $ rclone --config local.conf copy 1.json local:data/sample/ schema, table 생성 default schema (db) 생성 및 external_location table 생성 -- hive 의 schema(db) 생성 create schema hive.default; -- table 생성 create table hive.default.sample ( id varchar, name varchar ) with ( format = 'json', external_location = 's3a://data/sample/' ); -- select select * from sample; 쿼리 테스트 (dbeaver)