Question

我有一张cpu利用率数据表。每行都有一个主机名，利用率百分比和时间戳。我想为特定主机选择所有主机和cpu利用率最高时间戳。

下面是我的表结构，插入和选择

CREATE TABLE cpu_utilization
(
  id integer NOT NULL,
  hostname character varying(255),
  "timestamp" bigint,
  cpuutilizationpercentage integer,
  CONSTRAINT cpu_utilization_pkey PRIMARY KEY (id)
)
WITH (
  OIDS=FALSE
);
ALTER TABLE cpu_utilization
  OWNER TO postgres;


INSERT INTO cpu_utilization
VALUES (1,'host1','111111',10);
INSERT INTO cpu_utilization
VALUES (2,'host1','111112',12);
INSERT INTO cpu_utilization
VALUES (3,'host1','111113',15);

INSERT INTO cpu_utilization
VALUES (4,'host2','111111',20);
INSERT INTO cpu_utilization
VALUES (5,'host2','111112',22);
INSERT INTO cpu_utilization
VALUES (6,'host2','111113',25);

INSERT INTO cpu_utilization
VALUES (7,'host3','111111',10);

我正在执行此选择查询

select * from cpu_utilization where timestamp in ( select max(timestamp) from cpu_utilization group by hostname );

我得到的输出是 -

id; hostname    timestamp   cpuutilizationpercentage
1;  "host1";    111111;     10
3;  "host1";    111113;     15
4;  "host2";    111111;     20
6;  "host2";    111113;     25
7;  "host3";    111111;     10

但我期待输出为 -

id; hostname    timestamp   cpuutilizationpercentage
3;  "host1";    111113;     15
6;  "host2";    111113;     25
7;  "host3";    111111;     10

我的选择查询有什么问题，任何想法？

Answer 1

感谢@cha，我错误地在id上执行where子句，而不是它应该在主机名上。我用 -

获得了所需的输出

select * from cpu_utilization where timestamp = (select max(timestamp) from cpu_utilization cpuutil where cpuutil.hostname=cpu_utilization.hostname);

根据最高时间戳选择每个主机名的利用率百分比

1 个答案: