數(shù)字函數(shù)
- abs
求絕對值
select abs(-20.5); -- 20.5
- ceil
取上限
select ceil(14.2); -- 15
- floor
取下限
select floor(14.8); -- 14
- least
取最小值
select least(11, 4, 5);
- mod
取余數(shù)
select mod(9, 4); -- 1
- round
四舍五入
select round(11.2); -- 11
select round(11.8); -- 12
聚集函數(shù)
數(shù)據(jù)準備:
create table table9(
id int4,
num int4
);
insert into table9 (id, num) values (1, 12);
insert into table9 (id, num) values (1, 15);
insert into table9 (id, num) values (2, 11);
insert into table9 (id, num) values (3, 17);
insert into table9 (id, num) values (3, 14);
- count
統(tǒng)計行數(shù)
select count(*) from table9; -- 5
- avg
取均值
select id, avg(num) from table9 group by id;
- max
取最大值
select id, max(num) from table9 group by id;
- min
取最小值
select id, min(num) from table9 group by id;
- sum
求和
select id, sum(num) from table9 group by id;
字符函數(shù)
數(shù)據(jù)準備:
create table table8(
id int4,
name text,
line text
);
insert into table8 (id, name, line) values (1001, 'aladdin', 'xAddcwWf345ww');
- char_length
字符數(shù)
select char_length(line) from table8; -- 13
- concat
拼接字符串
select concat(id, name, '&&') from table8; -- 1001aladdin&&
- concat_ws
拼接字符串,并指定分隔符
select concat_ws('_', id, name, line) from table8; -- 1001_aladdin_xAddcwWf345ww
- substring
截取字符串
-- 從 3 開始,截取 2 個
select substring(line, 3, 2) from table8;
- split_part
分割字符串
select split_part('aladdin.im', '.', 1); -- aladdin
- upper
將字符串轉(zhuǎn)大寫
select upper(line) from table8;
- lower
轉(zhuǎn)小寫
select lower(line) from table8; -- xaddcwwf345ww
- ltrim
去除字符串左邊空格
select ltrim(' xxdd ');
- rtrim
去除字符串右邊空格
select rtrim(' xx ');
- position
字符串首次出現(xiàn)的位置
select position('Add' in line) from table8;
- repeat
重復字符串
select repeat(name, 3) from table8; -- aladdinaladdinaladdin
- replace
替換字符串
-- 將所有w替換成*
select replace(line, 'w', '*') from table8; -- xAddc*Wf345**
- left
返回最左面 n 個字符
select left(line, 5) from table8;
- right
返回最右面 n 個字符
select right(line, 3) from table8; -- 5ww
- rpad
字符不滿,用 * 補全
-- 字符不滿 10 個,用 # 補滿
select rpad(name, 10, '#') from table8;
日期函數(shù)
- extract
select extract(year from now()); -- 2019
-- week 一年中第幾周
-- doy 一年中第幾天
- 四則運算
select now() + interval '1 year'; -- 2020-04-10 09:43:35...
- age
時間間隔
select age('2018-5-23'::date, '2015-5-23'::date); -- 3 years 0 mons 0 days 0 hours 0 mins 0.00 secs
聚合函數(shù)
- string_agg
將結(jié)果集的某個字段所有行連接成字符串
-- 第一個參數(shù)為合并字段,第二個參數(shù)為分隔符
-- 可以用于分組合并
select id, string_agg(name, ',') from table5 group by id;
- array_agg
與 string_agg 類似,但是它返回的是一個數(shù)組
select id, array_agg(name) from table5 group by id;
- array_to_string
這個函數(shù)可以將數(shù)組類型數(shù)據(jù)轉(zhuǎn)換成字符串
select id, array_to_string(array_agg(name), ',') from table5 group by id;
- array_to_json
與 array_to_string 類似,這個函數(shù)只是將數(shù)組轉(zhuǎn)換成了 Json
select array_to_json(arr_int) from table6;
轉(zhuǎn)化函數(shù)
- translate
將字符串中的內(nèi)容進行轉(zhuǎn)化
create table public.table3(
msg text
);
insert into public.table3(msg) values ('ALADDIN 2019');
-- translate 函數(shù)
select
translate(msg, '0123456789ABCD', '##########%%%%')
from table3; -- %L%%%IN ####
- replace
將字符串中的字段進行轉(zhuǎn)化
select
replace(msg, 'ALADDIN', 'HELLO')
from table3; -- HELLO 2019
- coalesce
將 null 值轉(zhuǎn)化為特定值
-- 將 null -> 0
select coalesce(t2.sal, 0) from table2 t2;
窗口函數(shù)
窗口函數(shù)不會將結(jié)果集進行分組計算輸出一行,而是將計算后的結(jié)果集輸出到結(jié)果集上,可以簡化 SQL 代碼。
一般情況下,窗口函數(shù)能做的事情,都是可以通過復雜的 SQL 語句來實現(xiàn)。
- avg()
用于計算分組后的平均值 ( avg() 為聚合函數(shù) + over() 用來進行分組操作 = 分組求均值 )
create table student_score(
stu_id int4,
subject text,
score double precision
);
insert into student_info(id, name) values (1001, 'aladdin'), (1002, 'bilib'), (1003, 'chrome');
insert into
student_score(stu_id, subject, score)
values
(1001, 'chinese', 90.0), (1001, 'math', 100.0), (1001, 'english', 80.0),
(1002, 'chinese', 100.0), (1002, 'math', 80.0), (1002, 'english', 60.0),
(1003, 'chinese', 60), (1003, 'math', 90), (1003, 'english', 100.0);
-- 統(tǒng)計學個各科成績和各科的平均成績
select
ss.subject, si.name, ss.score, avg(ss.score) over(partition by ss.subject)
from
student_info si left join student_score ss on si.id = ss.stu_id;
- row_number()
為分組后數(shù)據(jù)標注行號
select
si.*, ss.subject, ss.score, row_number() over (partition by ss.subject order by ss.score)
from
student_info si left join student_score ss on si.id = ss.stu_id;
- rank() 和 dense_rank()
rank() 函數(shù)與 row_number() 類似,只是 rank() 會考慮并列的問題,rank() 和 dense_rank() 的區(qū)別是,rank() 會產(chǎn)生排序間隙,dense_rank() 不會產(chǎn)生排序間隙:
rank() 的排序結(jié)果可能是:1, 1, 3, 4, 4, 6
dense_rank() 的排序結(jié)果是:1, 1, 2, 3, 3, 4
- lag()
將某字段的數(shù)據(jù)進行偏移拼接:
select
si.*, ss.subject, ss.score, lag(ss.score, 1, '100.0') over(partition by ss.subject order by score)
-- lag(field, offset, default value),如果向上偏 offset = -X
from
student_info si left join student_score ss on si.id = ss.stu_id;
- first_value() & last_value()
取結(jié)果集分組后的第一條 / 最后一條數(shù)據(jù)
select
ss.subject ,first_value(ss.score) over (partition by subject order by score)
from
student_score ss;
- nth_value()
用來取結(jié)果集某一指定行的數(shù)據(jù)
select
ss.subject ,nth_value(ss.score, 2) over (partition by subject order by score)
from
student_score ss;