如何做PostgreSQL交叉表查询列中缺少值

Question

我正在尝试使用PostgreSQL 9.5交叉表查询创建转向时间报告,其中引用被分配到第1,2,3,4,> 4天(参见下面的输出).我有查询工作但是,如果我运行查询并且缺少第2天的值,则整个行会向左移动一个单元格.所以第二天持有第3天的价值,第3天持有第4天的价值等等......有人可以帮助我如何保持这个例子中的第2天是空白还是零而不是将行向左移动？

提前致谢.

脚本:

DROP TABLE IF EXISTS dt_temp;
CREATE TABLE dt_temp(id SERIAL, day int , referrals bigint);

INSERT INTO dt_temp(day, referrals) VALUES(1, 60);
INSERT INTO dt_temp(day, referrals) VALUES(2, 15);
INSERT INTO dt_temp(day, referrals) VALUES (3, 13);
INSERT INTO dt_temp(day, referrals) VALUES (4, 10);
INSERT INTO dt_temp(day, referrals) VALUES (5, 1);
INSERT INTO dt_temp(day, referrals) VALUES (6, 2);
INSERT INTO dt_temp(day, referrals) VALUES (7, 1);
INSERT INTO dt_temp(day, referrals) VALUES (8, 1);

Select * from crosstab(
    $$
    Select 'INDICATOR1' "INDICATOR", days::text, sum(referrals)::text     
from (
      SELECT CASE  
      WHEN day > 4 THEN '>4'
      ELSE day::text
      END "days",
   referrals
   FROM dt_temp
   ) "t"
group by 1,2
order by 2
$$
) AS dx2ref(Indicator text, Day1 text,  Day2 text, Day3 text, Day4 text, "Day > 4" text )

OUTPUT:

indicator  Day1 Day2 Day3 Day4 Day > 4
--------------------------------------
INDICATOR1 60   15   13   10   5

Answer 1

使用函数的第二种形式crosstab(text source_sql, text category_sql):

select * from crosstab(
    $$
        select 'INDICATOR1' "INDICATOR", days::text, sum(referrals)::text     
        from (
            select case  
                when day > 4 then '>4'
                else day::text
            end "days",
            referrals
            from dt_temp
            ) "t"
        group by 1,2
        order by 2
    $$,
    $$
        select days from (values ('1'), ('2'), ('3'), ('4'), ('>4')) t(days)
    $$
) AS dx2ref(Indicator text, Day1 text,  Day2 text, Day3 text, Day4 text, "Day > 4" text )


 indicator  | day1 | day2 | day3 | day4 | Day > 4 
------------+------+------+------+------+---------
 INDICATOR1 | 60   | 15   | 13   | 10   | 5
(1 row)