云服务器内容精选

  • 分析语句语法 完整的分析语句语法如下: SELECT [DISTINCT] (* | expression) [AS alias] [, ...] [GROUP BY expression [, ...] [HAVING predicates]] [ORDER BY expression [ASC | DESC] [, ...]] [LIMIT size OFFSET offset] 父主题: 查询与分析语法-SQL语法
  • 多个子查询 with t1 as(select name,max(price) as maxprice from fruit group by name), t2 as(select name,avg(price) as avgprice from fruit group by name) select t1.*,t2.* from t1 join t2 on t1.name = t2.name;
  • 使用示例 创建一个名为table1的普通表: CREATE TABLE example_db.table1 ( k1 TINYINT, k2 DECIMAL(10, 2) DEFAULT "10.5", k3 CHAR(10) COMMENT "string column", k4 INT NOT NULL DEFAULT "1" COMMENT "int column" ) COMMENT "table comment" DISTRIBUTED BY HASH(k1) BUCKETS 32; 创建一个名为table2的分区表。 使用event_day列作为分区列,建立3个分区:p201706、p201707、p201708,取值为: p201706:范围为[最小值, 2017-07-01) p201707:范围为[2017-07-01, 2017-08-01) p201708:范围为[2017-08-01, 2017-09-01) 每个分区使用siteid进行哈希分桶,桶数为10。 创建表命令如下: CREATE TABLE table2 ( event_day DATE, siteid INT DEFAULT '10', citycode SMALLINT, username VARCHAR(32) DEFAULT '', pv BIGINT SUM DEFAULT '0' ) AGGREGATE KEY(event_day, siteid, citycode, username) PARTITION BY RANGE(event_day) ( PARTITION p201706 VALUES LESS THAN ('2017-07-01'), PARTITION p201707 VALUES LESS THAN ('2017-08-01'), PARTITION p201708 VALUES LESS THAN ('2017-09-01') ) DISTRIBUTED BY HASH(siteid) BUCKETS 10 PROPERTIES("replication_num" = "2"); Doris创建表时限制至少指定2副本,以保证高可用。 可以对Table增加上卷表(Rollup)以提高查询性能。 表的列的Null属性默认为true,会对查询性能有一定的影响。 Doris表必须指定分桶列。 查看表内容: SHOW TABLES; +----------------------+ | Tables_in_example_db | +----------------------+ | table1 | | table2 | +----------------------+ 2 rows in set (0.01 sec) DESC table1; +----------+-------------+------+-------+---------+-------+ | Field | Type | Null | Key | Default | Extra | +----------+-------------+------+-------+---------+-------+ | siteid | int(11) | Yes | true | 10 | | | citycode | smallint(6) | Yes | true | N/A | | | username | varchar(32) | Yes | true | | | | pv | bigint(20) | Yes | false | 0 | SUM | +----------+-------------+------+-------+---------+-------+ 4 rows in set (0.00 sec) DESC table2; +-----------+-------------+------+-------+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-----------+-------------+------+-------+---------+-------+ | event_day | date | Yes | true | N/A | | | siteid | int(11) | Yes | true | 10 | | | citycode | smallint(6) | Yes | true | N/A | | | username | varchar(32) | Yes | true | | | | pv | bigint(20) | Yes | false | 0 | SUM | +-----------+-------------+------+-------+---------+-------+ 5 rows in set (0.00 sec)
  • 基本语法 CREATE TABLE [IF NOT EXISTS] [database.]table ( column_definition_list, [index_definition_list] ) [engine_type] [keys_type] [table_comment] [partition_info] distribution_desc [rollup_list] [properties] [extra_properties]
  • 语法格式 查询数据 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 [ WITH [ RECURSIVE ] with_query [, ...] ] SELECT [/*+ plan_hint */] [ ALL | DISTINCT [ ON ( expression [, ...] ) ] ] { * | {expression [ [ AS ] output_name ]} [, ...] } [ FROM from_item [, ...] ] [ WHERE condition ] [ GROUP BY grouping_element [, ...] ] [ HAVING condition [, ...] ] [ WINDOW {window_name AS ( window_definition )} [, ...] ] [ { UNION | INTERSECT | EXCEPT | MINUS } [ ALL | DISTINCT ] select ] [ ORDER BY {expression [ [ ASC | DESC | USING operator ] | nlssort_expression_clause ] [ NULLS { FIRST | LAST } ]} [, ...] ] [ LIMIT { [offset,] count | ALL } ] [ OFFSET start [ ROW | ROWS ] ] [ FETCH { FIRST | NEXT } [ count ] { ROW | ROWS } ONLY ] [ {FOR { UPDATE | SHARE } [ OF table_name [, ...] ] [ NOWAIT | WAIT N]} [...] ] TABLE { ONLY { (table_name) | table_name } | table_name [ * ]}; condition和expression中可以使用targetlist中表达式的别名。 只能同一层引用。 只能引用targetlist中的别名。 只能是后面的表达式引用前面的表达式。 不能包含volatile函数。 不能包含Window function函数。 不支持在join on条件中引用别名。 targetlist中有多个要应用的别名则报错。 其中子查询with_query为: 1 2 with_query_name [ ( column_name [, ...] ) ] AS [ [ NOT ] MATERIALIZED ] ( {select | values | insert | update | delete} ) 其中指定查询源from_item为: 1 2 3 4 5 6 7 {[ ONLY ] table_name [ * ] [ partition_clause ] [ [ AS ] alias [ ( column_alias [, ...] ) ] ] [ TABLESAMPLE sampling_method ( argument [, ...] ) [ REPEATABLE ( seed ) ] ] |( select ) [ AS ] alias [ ( column_alias [, ...] ) ] |with_query_name [ [ AS ] alias [ ( column_alias [, ...] ) ] ] |function_name ( [ argument [, ...] ] ) [ AS ] alias [ ( column_alias [, ...] | column_definition [, ...] ) ] |function_name ( [ argument [, ...] ] ) AS ( column_definition [, ...] ) |from_item [ NATURAL ] join_type from_item [ ON join_condition | USING ( join_column [, ...] ) ]} 其中group子句为: 1 2 3 4 5 6 ( ) | expression | ( expression [, ...] ) | ROLLUP ( { expression | ( expression [, ...] ) } [, ...] ) | CUBE ( { expression | ( expression [, ...] ) } [, ...] ) | GROUPING SETS ( grouping_element [, ...] ) 其中指定分区partition_clause为: 1 2 PARTITION { ( partition_name ) | FOR ( partition_value [, ...] ) } 指定分区只适合普通表。 其中设置排序方式nlssort_expression_clause为: 1 2 NLSSORT ( column_name, ' NLS_SORT = { SCHINESE_PINYIN_M | generic_m_ci } ' ) 其中,第二个参数可选generic_m_ci,仅支持纯英文不区分大小写排序。 简化版查询语法,功能相当于select * from table_name。 1 TABLE { ONLY {(table_name)| table_name} | table_name [ * ]};
  • 参数说明 model_name 对训练模型进行命名,模型名称具有唯一性约束。 取值范围:字符串,需要符合标识符的命名规范。 architecture_name 训练模型的算法类型。 取值范围:字符型,当前支持:logistic_regression、linear_regression、svm_classification、kmeans。 attribute_list 枚举训练模型的输入列名。 取值范围:字符型,需要符合数据属性名的命名规范。 attribute_name 在监督学习任务中训练模型的目标列名(可进行简单的表达式处理)。 取值范围:字符型,需要符合数据属性名的命名规范。 subquery 数据源。 取值范围:字符串,符合数据库SQL语法。
  • 示例 CREATE TABLE houses ( id INTEGER, tax INTEGER, bedroom INTEGER, bath DOUBLE PRECISION, price INTEGER, size INTEGER, lot INTEGER, mark text ); insert into houses(id, tax, bedroom, bath, price, size, lot, mark) VALUES (1,590,2,1,50000,770,22100,'a+'), (2,1050,3,2,85000,1410,12000,'a+'), (3,20,2,1,22500,1060,3500,'a-'), (4,870,2,2,90000,1300,17500,'a+'), (5,1320,3,2,133000,1500,30000,'a+'), (6,1350,2,1,90500,850,25700,'a-'), (7,2790,3,2.5,260000,2130,25000,'a+'), (8,680,2,1,142500,1170,22000,'a-'), (9,1840,3,2,160000,1500,19000,'a+'), (10,3680,4,2,240000,2790,20000,'a-'), (11,1660,3,1,87000,1030,17500,'a+'), (12,1620,3,2,118500,1250,20000,'a-'), (13,3100,3,2,140000,1760,38000,'a+'), (14,2090,2,3,148000,1550,14000,'a-'), (15,650,3,1.5,65000,1450,12000,'a-'); CREATE MODEL price_model USING logistic_regression FEATURES size, lot TARGET mark FROM HOUSES WITH learning_rate=0.88, max_iterations=default;
  • 语法格式 CREATE MODEL model_name USING algorithm_name [FEATURES { {expression [ [ AS ] output_name ]} [, ...] }] [TARGET { {expression [ [ AS ] output_name ]} [, ...] }] FROM { table_name | select_query } WITH hyperparameter_name = { hyperparameter_value | DEFAULT } [, ...] }
  • 示例 START TRANSACTION; START TRANSACTION ISOLATION LEVEL REPEATABLE READ; START TRANSACTION READ WRITE; START TRANSACTION ISOLATION LEVEL READ COMMITTED, READ ONLY; START TRANSACTION READ WRITE, ISOLATION LEVEL SERIALIZABLE; 不支持嵌套事务,也就是开启事务后,在commit之前不能再开启其他事务。
  • 描述 向表中插入新的数据行。 如果指定了列名列表,那么这些列名列表必须与query语句产生列列表名完全匹配。表中不在列名列表中的每一列,其值会设置为null。 如果没有指定列名列表,则query语句产生的列必须与将要插入的列完全匹配。 使用insert into时,会往表中追加数据,而使用insert overwrite时,如果表属性“auto.purge”被设置为“true”,直接删除原表数据,再写入新的数据。 如果对象表是分区表时,insert overwrite会删除对应分区的数据而非所有数据。 insert into后面的table关键字为可选,以兼容hive语法。
  • 示例 创建fruit和fruit_copy表: create table fruit (name varchar,price double); create table fruit_copy (name varchar,price double); 向fruit表中插入一行数据: insert into fruit values('LIchee',32); -- 兼容写法示例,带上table关键字 insert into table fruit values('Cherry',88); 向fruit表中插入多行数据: insert into fruit values('banana',10),('peach',6),('lemon',12),('apple',7); 将fruit表中的数据行加载到fruit_copy表中,执行后表中有5条记录: insert into fruit_copy select * from fruit; 先清空fruit_copy表,再将fruit中的数据加载到表中,执行之后表中有2条记录: insert overwrite fruit_copy select * from fruit limit 2; 对于varchar类型,仅当目标表定义的列字段长度大于源表的实际字段长度时,才可以使用INSERT... SELECT...的形式从源表中查数据并且插入到目标表: create table varchar50(c1 varchar(50)); insert into varchar50 values('hetuEngine'); create table varchar100(c1 varchar(100)); insert into varchar100 select * from varchar50; 分区表使用insert overwrite语句时,只会清理插入值所在分区的数据,而不是整个表: --创建表 create table test_part (id int, alias varchar) partitioned by (dept_id int, status varchar); insert into test_part partition(dept_id=10, status='good') values (1, 'xyz'), (2, 'abc'); select * from test_part order by id; id | alias | dept_id | status ----|-------|---------|-------- 1 | xyz | 10 | good 2 | abc | 10 | good (2 rows) --清理分区partition(dept_id=25, status='overwrite'),并插入一条数据 insert overwrite test_part (id, alias, dept_id, status) values (3, 'uvw', 25, 'overwrite'); select * from test_part ; id | alias | dept_id | status ----|-------|---------|----------- 1 | xyz | 10 | good 2 | abc | 10 | good 3 | uvw | 25 | overwrite --清理分区partition(dept_id=10, status='good'),并插入一条数据 insert overwrite test_part (id, alias, dept_id, status) values (4, 'new', 10, 'good'); select * from test_part ordr; id | alias | dept_id | status ----|-------|---------|----------- 3 | uvw | 25 | overwrite 4 | new | 10 | good (2 rows) --分区表插入数据 create table test_p_1(name string, age int) partitioned by (provice string, city string); create table test_p_2(name string, age int) partitioned by (provice string, city string); -- 填充数据到test_p_1 insert into test_p_1 partition (provice = 'hebei', city= 'baoding') values ('xiaobei',15),( 'xiaoming',22); -- 根据test_p_1 插入数据到test_p_2 -- 方式一 from test_p_1 insert into table test_p_2 partition (provice = 'hebei', city= 'baoding') select name,age; -- 方式二 insert into test_p_2 partition(provice = 'hebei', city= 'baoding') select name,age from test_p_1;
  • 语法 INSERT { INTO | OVERWRITE } [TABLE] table_name [(column_list)] [ PARTITION (partition_clause)] {select_statement | VALUES (value [, value ...]) [, (value [, value ...]) ...] } FROM from_statement INSERT OVERWRITE TABLE tablename1 [PARTITION (partcol1=val1, partcol2=val2 ...)] select_statement FROM from_statement INSERT INTO TABLE tablename1 [PARTITION (partcol1=val1, partcol2=val2 ...) select_statement
  • 限制 如果数据表中只有一个字段,且字段类型为row、struct,那么插入数据时需要用row对类型进行包裹。 -- 单字段表插入复杂类型需要用row()包裹 CREATE TABLE test_row (id row(c1 int, c2 string)); INSERT INTO test_row values row(row(1, 'test')); --多字段表复杂类型可以直接插入 CREATE TABLE test_multy_value(id int, col row(c1 int, c2 string)); INSERT INTO test_multy_value values (1,row(1,'test'));
  • Grouping 语法简介: 当group by语句带with rollup/cube选项时,Grouping才有意义。 CUBE生成的结果集显示了所选列中值的所有组合的聚合。 ROLLUP生成的结果集显示了所选列中值的某一层次结构的聚合。 Grouping:当用CUBE或ROLLUP运算符添加行时,附加的列输出值为1;当所添加的行不是由CUBE或ROLLUP产生时,附加列值为0。 例如,Hive中有一张表“table_test”,表结构如下所示: +----------------+-------------------+--+ | table_test.id | table_test.value | +----------------+-------------------+--+ | 1 | 10 | | 1 | 15 | | 2 | 20 | | 2 | 5 | | 2 | 13 | +----------------+-------------------+--+ 执行如下语句: select id,grouping(id),sum(value) from table_test group by id with rollup; 得到如下结果: +-------+-----------------+------+--+ | id | groupingresult | sum | +-------+-----------------+------+--+ | 1 | 0 | 25 | | NULL | 1 | 63 | | 2 | 0 | 38 | +-------+-----------------+------+--+
  • EXCEPT、INTERSECT EXCEPT返回两个结果集的差(即从左查询中返回右查询没有找到的所有非重复值)。 INTERSECT返回两个结果集的交集(即两个查询都返回的所有非重复值)。 例如,Hive中有两张表“test_table1”、“test_table2”。 “test_table1”表结构如下所示: +-----------------+--+ | test_table1.id | +-----------------+--+ | 1 | | 2 | | 3 | | 4 | +-----------------+--+ “test_table2”表结构如下所示: +-----------------+--+ | test_table2.id | +-----------------+--+ | 2 | | 3 | | 4 | | 5 | +-----------------+--+ 执行如下的EXCEPT语句: select id from test_table1 except select id from test_table2; 显示如下结果: +--------------+--+ | _alias_0.id | +--------------+--+ | 1 | +--------------+--+ 执行INTERSECT语句: select id from test_table1 intersect select id from test_table2; 显示如下结果: +--------------+--+ | _alias_0.id | +--------------+--+ | 2 | | 3 | | 4 | +--------------+--+