Question

我从表中获得了一些具有预期和实际结果值的统计数据。

这句话完美无缺：

select model_ref AS [Model],
     count(id) AS Tested,
     count(case when expected_result = predicted_result then predicted_result end) AS OK,
     count(case when NOT(expected_result = predicted_result) then expected_result end) AS Wrong,
     ( (count(case when expected_result = predicted_result then predicted_result end)*1.0) / count(id) * 100  ) AS [Accuracy (%)]
from Results

但这不是：

select model_ref AS [Model],
     count(id) AS Tested,
     count(case when expected_result = predicted_result then predicted_result end) AS OK,
     count(case when NOT(expected_result = predicted_result) then expected_result end) AS Wrong,
     ( OK *1.0) / Tested * 100  ) AS [Accuracy (%)]
from Results

区别仅在于，我想通过使用先前定义的列名来缩短表达式 - 已测试和确定

有可能吗？

P.S .: 如果它具有任何重要性，则表格为：

CREATE TABLE Results(
id INTEGER PRIMARY KEY AUTOINCREMENT,
expected_result TEXT,
image_id TEXT,
model_ref TEXT,
predicted_result TEXT,
test_date TEXT)

Answer 1

在大多数数据库中，别名不能在查询的同一“级别”（或范围）上使用（因为在解析查询时可能无法解析别名）。你可以做的是将查询包装在一个外部查询中，该查询可以访问这样的别名：

select 
    Model,
    Tested,
    OK,
    Wrong,
    (( OK *1.0) / Tested * 100  ) AS [Accuracy (%)] 
from (
    select model_ref AS [Model],
       count(id) AS Tested,
       count(case when expected_result = predicted_result then predicted_result end) AS OK,
       count(case when NOT(expected_result = predicted_result) then expected_result end) AS Wrong
    from Results
    group by model_ref
) a

使用先前在同一SQL语句中定义的列短名称

1 个答案: