正如@JohnCappelletti在评论中所说，你可以像这样使用漂亮的Sign函数（假设值〉=0）：

SELECT 
, P.January as 'Jan'
, P.February as 'Feb'
, P.March as 'Mar'
, Sign(P.January) + Sign(P.February) + Sign(P.March) /*other months*/ as Count
FROM P

如果没有负值，我们可以利用CONVERT(bit, 13)产生1的事实，并从CROSS APPLY中获得SUM：

SELECT s.Jan, s.Feb, s.Mar, 
  [Count] = SUM(CONVERT(int, CONVERT(bit, x.v)))
FROM dbo.YourTableName AS s
CROSS APPLY (VALUES(s.Jan),(s.Feb),(s.Mar)) AS x(v)
GROUP BY s.Jan, s.Feb, s.Mar;

• 示例db<>fiddle

如果有不应计数的负值，只需添加col + ABS(col)即可将负值更改为0，将正值加倍，并保持0不变。

SELECT s.Jan, s.Feb, s.Mar, 
  [Count] = SUM(CONVERT(int, CONVERT(bit, x.v + ABS(x.v))))
  -------------------------------------------^^^^^^^^^^^
FROM dbo.YourTableName AS s
CROSS APPLY (VALUES(s.Jan),(s.Feb),(s.Mar)) AS x(v)
GROUP BY s.Jan, s.Feb, s.Mar;

下面是一个灵活的解决方案，无论列数如何，它都能正常工作。

SQL语言

-- DDL and sample data population, start
DECLARE @tbl TABLE (Jan int, Feb int, Mar INT, Apr INT, May INT);
INSERT @tbl (Jan,Feb,Mar, Apr, May) VALUES
(13, 0,  25, 0, 2),
(11, 10, 4,  1, 18),
(0,  0,  7,  0, 7);
-- DDL and sample data population, end

SELECT t.*
   , [Count] = x.value('count(/root/*[not(text()="0")]/text())', 'INT')
FROM @tbl AS t
CROSS APPLY (SELECT t.* FOR XML PATH(''), TYPE, ROOT('root')) AS t1(x);

产出

| 一月|二月|马尔|四月|五月|计数|
| - ------|- ------|- ------|- ------|- ------|- ------|
| 十三|无|二十五|无|第二章|三个|
| 十一|十个|四个|1个|十八|五个|
| 无|无|七|无|七|第二章|

我认为这是IIF的一个很好的用例，请参见documentation。
查询将是这样一个：

SELECT 
Jan, Feb, Mar,
IIF(Jan > 0, 1, 0) + IIF(Feb > 0, 1, 0) + IIF(Mar > 0, 1, 0) AS [Count]
FROM p;