Question

我正在Azure上运行PostgreSQL 10.5。我想创建一个存储过程来返回一组行，所以我从安装脚本中复制所选数据的类型：

DROP TYPE t CASCADE;
CREATE TYPE t as (region varchar(24), county varchar(50), year int, value int);
DROP FUNCTION geo();
CREATE FUNCTION geo()
returns setof t language sql as $$
SELECT
region.name, county.name, tbl.year, tbl.value
FROM tbl
LEFT JOIN county ON tbl.lau1 = county.lau1
LEFT JOIN region ON county.regionid = region.regionid
$$;

我得到了错误：

ERROR:  return type mismatch in function declared to return t
DETAIL:  Final statement returns character instead of character varying at column 1.
CONTEXT:  SQL function "geo"

但是，类型与创建region和county表中的类型相同（并且我还包括了tbl的代码）：

CREATE TABLE county (

    LAU1    CHAR(9),
    NAME    VARCHAR(50) NOT NULL,
    REGIONID    INT NOT NULL,
    PRIMARY KEY(LAU1)
);

CREATE TABLE region (
    REGIONID    SMALLINT,
    NAME        VARCHAR(24) NOT NULL,
    PRIMARY KEY(REGIONID)
);

CREATE TABLE tbl (
    LAU1        CHAR(9),
    YEAR        INT,
    VALUE       INT,
    PRIMARY KEY(LAU1, YEAR)
);

三个表的示例：

=> select * from county limit 2;
   lau1    |   name    | regionid 
-----------+-----------+----------
 E06000015 | Derby     |        4
 E06000016 | Leicester |        4
(2 rows)

=> select * from region limit 2;
 regionid |           name           
----------+--------------------------
        1 | North East              
        2 | North West              
(2 rows)

=> select * from tbl limit 2;
   lau1    | year | value 
-----------+------+-------
 E06000047 | 1997 |  3241
 E06000005 | 1997 |   656
(2 rows)

我想在返回的值中保留与varchar的原始类型相同的类型。但是作为实验，如果我将元组的第一类型从varchar更改为char：

CREATE TYPE t as (region char(24), county varchar(50), year int, value int);

和其余代码一样，尽管第二列保留了varchar，它仍然可以工作！

select * from geo();
          region          |   county   | year | value 
--------------------------+------------+------+-------
 North East               | Hartlepool | 1997 |   625
(1 row)

怎么来？

最终语句返回字符而不是字符变化

0 个答案: