POSTGRESQL交叉表的实现

t_girl=# select * from score;
 name  | subject | score 
-------+---------+-------
 Lucy  | English |   100
 Lucy  | Physics |    90
 Lucy  | Math    |    85
 Lily  | English |    95
 Lily  | Physics |    81
 Lily  | Math    |    84
 David | English |   100
 David | Physics |    86
 David | Math    |    89
 Simon | English |    90
 Simon | Physics |    76
 Simon | Math    |    79
(12 rows)


Time: 2.066 ms

 name  | English | Physics | Math 
-------+---------+---------+------
 Simon |      90 |      76 |   79
 Lucy  |     100 |      90 |   85
 Lily  |      95 |      81 |   84
 David |     100 |      86 |   89

t_girl=# select name, 
t_girl-#  sum(case when subject = ‘English‘ then score else 0 end) as "English",
t_girl-#  sum(case when subject = ‘Physics‘ then  score else 0 end) as "Physics",
t_girl-#  sum(case when subject = ‘Math‘   then score else 0 end) as "Math" 
t_girl-#  from score
t_girl-#  group by name order by name desc;
 name  | English | Physics | Math 
-------+---------+---------+------
 Simon |      90 |      76 |   79
 Lucy  |     100 |      90 |   85
 Lily  |      95 |      81 |   84
 David |     100 |      86 |   89
(4 rows)


Time: 1.123 ms

t_girl=# SELECT *
FROM crosstab(‘select name,subject,score from score order by name desc‘,$$values (‘English‘::text),(‘Physics‘::text),(‘Math‘::text)$$)
AS score(name text, English int, Physics int, Math int);
 name  | english | physics | math 
-------+---------+---------+------
 Simon |      90 |      76 |   79
 Lucy  |     100 |      90 |   85
 Lily  |      95 |      81 |   84
 David |     100 |      86 |   89
(4 rows)


Time: 2.059 ms

t_girl=# select name,split_part(split_part(tmp,‘,‘,1),‘:‘,2) as "English",
t_girl-# split_part(split_part(tmp,‘,‘,2),‘:‘,2) as "Physics",
t_girl-# split_part(split_part(tmp,‘,‘,3),‘:‘,2) as "Math"
t_girl-# from
t_girl-# (
t_girl(# select name,string_agg(subject||‘:‘||score,‘,‘) as tmp from score group by name order by name desc
t_girl(# ) as T;
 name  | English | Physics | Math 
-------+---------+---------+------
 Simon | 90      | 76      | 79
 Lucy  | 100     | 90      | 85
 Lily  | 95      | 81      | 84
 David | 100     | 86      | 89
(4 rows)


Time: 2.396 ms

create or replace function func_ytt_crosstab_py ()
returns setof ytt_crosstab
as 
$ytt$
  for row in plpy.cursor("select name,string_agg(subject||‘:‘||score,‘,‘) as tmp from score group by name order by name desc"):
      a = row[‘tmp‘].split(‘,‘)
      yield (row[‘name‘],a[0].split(‘:‘)[1],a[1].split(‘:‘)[1],a[2].split(‘:‘)[1])
$ytt$ language plpythonu;


t_girl=# select name,english,physics,math from  func_ytt_crosstab_py();
 name  | english | physics | math 
-------+---------+---------+------
 Simon | 90      | 76      | 79
 Lucy  | 100     | 90      | 85
 Lily  | 95      | 81      | 84
 David | 100     | 86      | 89
(4 rows)


Time: 2.687 ms

5、 用PLPGSQL来实现

t_girl=# create type ytt_crosstab as (name text, English text, Physics text, Math text);
CREATE TYPE
Time: 22.518 ms


create or replace function func_ytt_crosstab ()
returns setof ytt_crosstab
as 
$ytt$
  declare v_name text := ‘‘;
                v_english text := ‘‘;
		v_physics text := ‘‘;
		v_math text := ‘‘;
		v_tmp_result text := ‘‘;
  declare cs1 cursor for select name,string_agg(subject||‘:‘||score,‘,‘) from score group by name order by name desc;
begin
  open cs1;
  loop
    fetch cs1 into v_name,v_tmp_result;
    exit when not found;
    v_english = split_part(split_part(v_tmp_result,‘,‘,1),‘:‘,2);
    v_physics = split_part(split_part(v_tmp_result,‘,‘,2),‘:‘,2);
    v_math = split_part(split_part(v_tmp_result,‘,‘,3),‘:‘,2);
    return query select v_name,v_english,v_physics,v_math;
  end loop;
end;
$ytt$ language plpgsql;


t_girl=# select name,English,Physics,Math from func_ytt_crosstab();
 name  | english | physics | math 
-------+---------+---------+------
 Simon | 90      | 76      | 79
 Lucy  | 100     | 90      | 85
 Lily  | 95      | 81      | 84
 David | 100     | 86      | 89
(4 rows)


Time: 2.127 ms