最近用presto引擎查數據，發現了語法和MYSQL，PG的稍許區別，寫此文章留念~~

presto是Facebook主持下運營的開源的分佈式SQL查詢引擎，用於針對各種大小（從千兆字節到千兆字節）的數據源運行交互式分析查詢。本文主要介紹常用SQL，具體可參考官方文檔：https://prestodb.github.io/docs/current/

1 數據類型

Boolean: true, false
Integer: tinyint, smallint, integer, bigint
Floating-Point: real, double
Fixed-Precision：DECIMAL
String：varchar, char, varbinary, json
Date and Time: date, time, time with time zone, timestamp, timestamp with time zone, interval year to month, interval day to second
Structural: array, map, row
Network Address: ipaddress
HyperLogLog: HyperLogLog, P4HyperLogLog
Quantile Digest: QDigest

2 SELECT 搜索查詢

[ WITH with_query [, ...] ]
SELECT [ ALL | DISTINCT ] select_expr [, ...]
[ FROM from_item [, ...] ]
[ WHERE condition ]
[ GROUP BY [ ALL | DISTINCT ] grouping_element [, ...] ]
[ HAVING condition]
[ { UNION | INTERSECT | EXCEPT } [ ALL | DISTINCT ] select ]
[ ORDER BY expression [ ASC | DESC ] [, ...] ]
[ LIMIT [ count | ALL ] ]

以下是這些參數可能的格式
- from_item
    table_name [ [ AS ] alias [ ( column_alias [, ...] ) ] ]
    from_item join_type from_item [ ON join_condition | USING ( join_column [, ...] ) ]

- join_type
    [ INNER ] JOIN
    LEFT [ OUTER ] JOIN
    RIGHT [ OUTER ] JOIN
    FULL [ OUTER ] JOIN
    CROSS JOIN
    
- grouping_element
    ()
    expression
    GROUPING SETS ( ( column [, ...] ) [, ...] )
    CUBE ( column [, ...] )
    ROLLUP ( column [, ...] )

2.1 with 子句

with 定義要在查詢中使用的命名關係

WITH x AS (SELECT a, MAX(b) AS b FROM t GROUP BY a)
SELECT a, b FROM x;

等同於
SELECT a, b
FROM (
  SELECT a, MAX(b) AS b FROM t GROUP BY a
) AS x;


也可以用於多條定義
WITH
  t1 AS (SELECT a, MAX(b) AS b FROM x GROUP BY a),
  t2 AS (SELECT a, AVG(d) AS d FROM y GROUP BY a)
SELECT t1.*, t2.*
FROM t1
JOIN t2 ON t1.a = t2.a;


也可以鏈式使用
WITH
  x AS (SELECT a FROM t),
  y AS (SELECT a AS b FROM x),
  z AS (SELECT b AS c FROM y)
SELECT c FROM z;

2.2 GROUP BY 子句

2.2.1 GROUP BY

當在select語句中使用group by時，所有輸出表達式都必須是聚合函數或group by子句中存在的列。

按字段nationkey分組，並查出各組數量，以下兩種寫法是一致的，by 2 代表以輸出表達式第2列做分組
SELECT count(*), nationkey FROM customer GROUP BY 2;
SELECT count(*), nationkey FROM customer GROUP BY nationkey;

也可以不輸出指定分組的列，如下
SELECT count(*) FROM customer GROUP BY mktsegment;

2.2.2 GROUPING SETS

可以指定多個列進行分組，結果列中不屬於分組列的將被設置爲NUll。
具有複雜分組語法(GROUPING SETS, CUBE 或 ROLLUP)的查詢只從基礎數據源讀取一次，而使用UNION ALL的查詢將讀取基礎數據三次。這就是當數據源不具有確定性時，使用UNION ALL的查詢可能會產生不一致的結果的原因。

有一個表：SELECT * FROM shipping;

origin_state | origin_zip | destination_state | destination_zip | package_weight
--------------+------------+-------------------+-----------------+----------------
 California   |      94131 | New Jersey        |            8648 |             13
 California   |      94131 | New Jersey        |            8540 |             42
 New Jersey   |       7081 | Connecticut       |            6708 |            225
 California   |      90210 | Connecticut       |            6927 |           1337
 California   |      94131 | Colorado          |           80302 |              5
 New York     |      10002 | New Jersey        |            8540 |              3
(6 rows)

SELECT origin_state, origin_zip, destination_state, sum(package_weight)
FROM shipping
GROUP BY GROUPING SETS (
    (origin_state),
    (origin_state, origin_zip),
    (destination_state));
    
這個的查詢在邏輯上等同於多個分組查詢的union all：

SELECT origin_state, NULL, NULL, sum(package_weight)
FROM shipping GROUP BY origin_state

UNION ALL

SELECT origin_state, origin_zip, NULL, sum(package_weight)
FROM shipping GROUP BY origin_state, origin_zip

UNION ALL

SELECT NULL, NULL, destination_state, sum(package_weight)
FROM shipping GROUP BY destination_state;


結果如下：
 origin_state | origin_zip | destination_state | _col0
--------------+------------+-------------------+-------
 New Jersey   | NULL       | NULL              |   225
 California   | NULL       | NULL              |  1397
 New York     | NULL       | NULL              |     3
 California   |      90210 | NULL              |  1337
 California   |      94131 | NULL              |    60
 New Jersey   |       7081 | NULL              |   225
 New York     |      10002 | NULL              |     3
 NULL         | NULL       | Colorado          |     5
 NULL         | NULL       | New Jersey        |    58
 NULL         | NULL       | Connecticut       |  1562
(10 rows)

2.2.3 CUBE

爲給定的列生成所有可能的分組，比如 (origin_state, destination_state) 的可能分組爲(origin_state, destination_state),
(origin_state),
(destination_state),
()

SELECT origin_state, destination_state, sum(package_weight)
FROM shipping
GROUP BY CUBE (origin_state, destination_state);

等同於

SELECT origin_state, destination_state, sum(package_weight)
FROM shipping
GROUP BY GROUPING SETS (
    (origin_state, destination_state),
    (origin_state),
    (destination_state),
    ());

2.2.4 ROLLUP

爲給定的列集生成部分可能的分類彙總

SELECT origin_state, origin_zip, sum(package_weight)
FROM shipping
GROUP BY ROLLUP (origin_state, origin_zip);

等同於

SELECT origin_state, origin_zip, sum(package_weight)
FROM shipping
GROUP BY GROUPING SETS ((origin_state, origin_zip), (origin_state), ());

2.2.5 group by, clue, rollup區別

比如按字段1，2，3來分組，group 只會聚合1，2，3分組，clue會展示所有層級分組，rollup只會展示1以下所有分組
用列表標識會更直觀

group by

1	2	3

clue

1	2	3
1	2
1
	2	3
	2
		3

rollup

1	2	3
1	2
1

2.2.6 group sets, clue, rollup 組合使用

SELECT origin_state, destination_state, origin_zip, sum(package_weight)
FROM shipping
GROUP BY
    GROUPING SETS ((origin_state, destination_state)),
    ROLLUP (origin_zip);

等同於

SELECT origin_state, destination_state, origin_zip, sum(package_weight)
FROM shipping
GROUP BY
    GROUPING SETS ((origin_state, destination_state)),
    GROUPING SETS ((origin_zip), ());
    
邏輯上等同於

SELECT origin_state, destination_state, origin_zip, sum(package_weight)
FROM shipping
GROUP BY GROUPING SETS (
    (origin_state, destination_state, origin_zip),
    (origin_state, destination_state));

2.3 HAVING

HAVING與聚合函數和GROUP BY一起使用，以過濾GROUP BY。

從customer表中選擇帳戶餘額大於指定值的組
SELECT count(*), mktsegment, nationkey,
       CAST(sum(acctbal) AS bigint) AS totalbal
FROM customer
GROUP BY mktsegment, nationkey
HAVING sum(acctbal) > 5700000
ORDER BY totalbal DESC;

2.4 UNION，INTERSECT， EXCEPT

query UNION [ALL | DISTINCT] query
query INTERSECT [DISTINCT] query
query EXCEPT [DISTINCT] query

- all: 最終結果集中包括所有行
- distinct: 組合結果集中只包含唯一的行
- 如果兩者都未指定，則行爲默認爲distinct。

區別
- intersect或except不支持all參數。
- 除非通過括號明確指定順序，否則將從左到右處理多個集合操作
- INTERSECT 優先級高於UNION和EXCEPT
    比如：
    A UNION B INTERSECT C EXCEPT D 
    等同於
    A UNION (B INTERSECT C) EXCEPT D

2.4.1 UNION

以下結果返回13和42
SELECT 13 UNION SELECT 42;

以下結果返回13和42
SELECT 13 UNION SELECT * FROM (VALUES 42, 13);

以下結果返回13，42 和 13
SELECT 13 UNION ALL SELECT * FROM (VALUES 42, 13);

2.4.2 INTERSECT

使用INTERSECT代表返回的最終結果集爲：INTERSECT之前的結果與INTERSECT查出的結果取交集

比如以下結果返回 13：
SELECT * FROM (VALUES 13, 42)
INTERSECT
SELECT 13;

2.4.3 EXCEPT

使用EXCEPT代表返回的最終結果集中排除EXCEPT查出的結果

比如以下結果返回 42：
SELECT * FROM (VALUES 13, 42)
EXCEPT
SELECT 13;

2.5 ORDER BY 排序

一般放到SELECT語句的最後，或在HAVING之前，默認ASC NULLS LAST，

ORDER BY expression [ ASC | DESC ] [ NULLS { FIRST | LAST } ] [, ...]

- ASC： 默認從小到大正序排序
- DESC： 倒序
- NULLS FIRST： NULL 值最大
- NULLS LAST： 默認 NULL 值最小

2.6 LIMIT

limit 5 代表只輸出5條結果，limit all 代表全部輸出，沒有數量限制

3 其他常用SQL

3.1 SCHEMA 操作

3.1.1 重命名 SCHEMA

ALTER SCHEMA name RENAME TO new_name

eg: 將 web 重命名爲 traffic
ALTER SCHEMA web RENAME TO traffic

3.1.2 創建 SCHEMA

CREATE SCHEMA [ IF NOT EXISTS ] schema_name
[ WITH ( property_name = expression [, ...] ) ]

- IF NOT EXISTS 比較安全，防止SCHEMA已存在的報錯
- WITH 可以給SCHEMA添加屬性，通過以下SQL可以查看所有屬性：
SELECT * FROM system.metadata.schema_properties

eg: 
創建一個名爲web的SCHEMA
CREATE SCHEMA web
創建一個在hive目錄下名爲sales的SCHEMA
CREATE SCHEMA hive.sales
如果名爲traffic的SCHEMA不存在那麼創建它
CREATE SCHEMA IF NOT EXISTS traffic

3.1.2 刪除 SCHEMA

DROP SCHEMA [ IF EXISTS ] schema_name

- IF EXISTS 可以防止SCHEMA不存在時的報錯

eg:
刪除名爲web的SCHEMA
DROP SCHEMA web
如果名爲web的SCHEMA存在，則刪除它
DROP SCHEMA IF EXISTS sales

3.2 TABLE 操作

3.2.1 創建 TABLE

CREATE TABLE [ IF NOT EXISTS ]
table_name (
  { column_name data_type [ COMMENT comment ] [ WITH ( property_name = expression [, ...] ) ]
  | LIKE existing_table_name [ { INCLUDING | EXCLUDING } PROPERTIES ] }
  [, ...]
)
[ COMMENT table_comment ]
[ WITH ( property_name = expression [, ...] ) ]


- IF NOT EXISTS 比較安全，防止TABLE已存在的報錯
- WITH 可以給TABLE添加屬性：
通過以下SQL可以查看所有表屬性
SELECT * FROM system.metadata.table_properties
通過以下SQL可以查看所有列屬性
SELECT * FROM system.metadata.column_properties
- COMMENT 爲表添加註釋
- LIKE 可用於在新表中包含來自現有表的所有列。可以指定多個LIKE子句，允許從多個表複製列。


eg:
創建一個名爲orders的表， 並添加表註釋
CREATE TABLE orders (
  orderkey bigint,
  orderstatus varchar,
  totalprice double,
  orderdate date
)
COMMENT 'A table to keep track of orders.'
WITH (format = 'ORC')

創建一個名爲bigger_orders的表，幷包含orders表的所有字段
CREATE TABLE bigger_orders (
  another_orderkey bigint,
  LIKE orders,
  another_orderdate date
)

3.2.2 查看建表語句

SHOW CREATE TABLE table_name

3.2.3 修改 TABLE

重命名
ALTER TABLE name RENAME TO new_name
添加字段
ALTER TABLE name ADD COLUMN column_name data_type [ COMMENT comment ] [ WITH ( property_name = expression [, ...] ) ]
刪除字段
ALTER TABLE name DROP COLUMN column_name
重命名字段
ALTER TABLE name RENAME COLUMN column_name TO new_column_name

3.2.4 刪除 TABLE

DROP TABLE  [ IF EXISTS ] table_name

- IF EXISTS 可以防止TABLE不存在時的報錯

eg:
刪除名爲web的TABLE
DROP TABLE web
如果名爲web的TABLE存在，則刪除它
DROP TABLE IF EXISTS sales

3.2.5 CREATE TABLE AS 使用搜索結果建新表

CREATE TABLE [ IF NOT EXISTS ] table_name [ ( column_alias, ... ) ]
[ COMMENT table_comment ]
[ WITH ( property_name = expression [, ...] ) ]
AS query
[ WITH [ NO ] DATA ]

eg:
創建一個新表orders_column_aliased，字段order_date, total_price分別來自於表orders的字段orderdate, totalprice
CREATE TABLE orders_column_aliased (order_date, total_price)
AS
SELECT orderdate, totalprice
FROM orders

3.3 ANALYZE 統計表和列信息

統計表和列信息，目前該語句只支持Hive connector。
ANALYZE table_name [ WITH ( property_name = expression [, ...] ) ]

- WITH 可以給查詢添加特定屬性：
通過以下SQL可以查看所有可以使用的屬性
SELECT * FROM system.metadata.analyze_properties

3.4 CALL 調用存儲過程

調用存儲過程，有些連接器，比如 PostgreSQL Connector,有自己的存儲過程，不能通過call調用
CALL procedure_name ( [ name => ] expression [, ...] )


eg:
傳入必傳參數，調用存儲過程
CALL test(123, 'apple');
傳入命名參數，調用存儲過程
CALL test(name => 'apple', id => 123);
不需要傳參，調用存儲過程
CALL catalog.schema.test();

3.5 START TRANSACTION，ROLLBACK，COMMIT 事務

開啓事務 （默認爲READ WRITE讀寫事務）
START TRANSACTION [ mode [, ...] ]
回滾事務
ROLLBACK [ WORK ]
提交事務
COMMIT [ WORK ]

- model 是以下的一種：
ISOLATION LEVEL { READ UNCOMMITTED | READ COMMITTED | REPEATABLE READ | SERIALIZABLE }
READ { ONLY | WRITE }

eg:
開始一個事務，默認爲READ WRITE讀寫事務
START TRANSACTION;
開始一個可重複讀事務
START TRANSACTION ISOLATION LEVEL REPEATABLE READ;
開始一個讀寫事務
START TRANSACTION READ WRITE;
開始一個提交讀/不可重複讀、只讀事務
START TRANSACTION ISOLATION LEVEL READ COMMITTED, READ ONLY;
開始一個讀寫串行化事務
START TRANSACTION READ WRITE, ISOLATION LEVEL SERIALIZABLE;

3.6 DELETE 刪除

有的連接器對於刪除有限制或者是不支持，需要看具體的連接器文檔

DELETE FROM table_name [ WHERE condition ]

eg:
刪除lineitem表裏的shipmode = 'AIR'的行
DELETE FROM lineitem WHERE shipmode = 'AIR';

刪除所有orders裏的數據
DELETE FROM orders;

3.8 PREPARE, EXECUTE, DEALLOCATE PREPARE

聲明一個名爲statement_name的SQL
PREPARE statement_name FROM statement
執行名爲statement_name的聲明
EXECUTE statement_name [ USING parameter1 [ , parameter2, ... ] ]
刪除名爲statement_name的聲明
DEALLOCATE PREPARE statement_name

eg:
準備一條sql語句
PREPARE my_select2 FROM
SELECT name FROM nation WHERE regionkey = ? and nationkey < ?;

執行這個語句，並加入？的參數
EXECUTE my_select2 USING 1, 3;

以上兩句相當於執行下面這條SQL：
SELECT name FROM nation WHERE regionkey = 1 AND nationkey < 3;

3.9 INSERT 插入數據

INSERT INTO table_name [ ( column [, ... ] ) ] query

eg：
往orders表裏插入數據，數據全部來源於new_orders表
INSERT INTO orders SELECT * FROM new_orders;

往cities表裏插入一條數據
INSERT INTO cities VALUES (1, 'San Francisco');

往cities表裏插入多條數據
INSERT INTO cities VALUES (2, 'San Jose'), (3, 'Oakland');

指定字段名往nation表裏插入多條數據，如果有字段未指定，則用字段默認值， 沒有默認值就是null
INSERT INTO nation (nationkey, name, regionkey, comment)
VALUES (26, 'POLAND', 3, 'no comment');

3.10 查看數據倉庫目錄，SCHEMA, TABLE，COLUMN

查看數據倉庫第一層目錄
SHOW CATALOGS [ LIKE pattern ]

查看所有SCHEMAS
SHOW SCHEMAS [ FROM catalog ] [ LIKE pattern ]

查看schema裏的表
SHOW TABLES [ FROM schema ] [ LIKE pattern ]

查看錶的字段類型，描述, 搜索出來的結果有：column type extra comment
DESCRIBE table_name
相當於 SHOW COLUMNS from table_name

presto 語法攻略

文章目錄