Informix窗口函数范围子句的行为类似于rows子句

Question

在Informix中考虑以下语句：

SELECT 
  a, 
  b,
  sum(a) over (order by b) "no frame",
  sum(a) over (order by b range between unbounded preceding and current row) "range",
  sum(a) over (order by b rows between unbounded preceding and current row) "rows"
FROM TABLE(SET{row(1, 1), row(2, 1), row(3, 2), row(4, 2)}) AS t(a, b);

令人惊讶的是，它为range和rows产生了相同的结果：

a |b |no frame |range |rows |
--|--|---------|------|-----|
1 |1 |3        |1     |1    |
2 |1 |3        |3     |3    |
3 |2 |10       |6     |6    |
4 |2 |10       |10    |10   |

这与我以前使用的任何RDBMS完全不同，包括PostgreSQL：

SELECT 
  a, 
  b,
  sum(a) over (order by b) "no frame",
  sum(a) over (order by b range between unbounded preceding and current row) "range",
  sum(a) over (order by b rows between unbounded preceding and current row) "rows"
FROM (values(1, 1), (2, 1), (3, 2), (4, 2)) AS t(a, b);

...产生预期的

a |b |no frame |range |rows |
--|--|---------|------|-----|
1 |1 |3        |3     |1    |
2 |1 |3        |3     |3    |
3 |2 |10       |10    |6    |
4 |2 |10       |10    |10   |

我错过了Informix和标准SQL之间的细微差别，还是一个bug？我正在使用official docker image，它对应于现在的Informix 12.10。

Answer 1

这是documented行为：

如果为窗口聚合函数指定ORDER子句但没有窗口框子句，则默认情况下，返回当前行和当前行之前的所有行，这等同于以下窗口框架规范：

无界前进与当前行之间的行程

我认为当没有窗口条款时，标准是RANGE。我很确定这是大多数其他数据库的默认设置。

Answer 2

作为一种解决方法，我可以使用0 PRECEDING（不应该允许，但很好）而不是CURRENT ROW。这产生了我预期的结果：

SELECT 
  a, 
  b,
  sum(a) over (order by b) "no frame",
  sum(a) over (order by b range between unbounded preceding and 0 preceding) "range",
  sum(a) over (order by b rows between unbounded preceding and current row) "rows"
FROM (values(1, 1), (2, 1), (3, 2), (4, 2)) AS t(a, b);

我现在得到：

a |b |no frame |range |rows |
--|--|---------|------|-----|
1 |1 |3        |3     |1    |
2 |1 |3        |3     |3    |
3 |2 |10       |10    |6    |
4 |2 |10       |10    |10   |