Question

我的字母数字文字如下

 ID Textfield
 1  estimated left ventricular ejection fraction 60-65%
 2  estimated left ventricular ejection fraction is 55-60%
 3  Left ventricular ejection fraction is approximately 40 to 50%
 4  Fractional Short 50 %( 28-48) LV mass 83 gm (<220) systolic function  left ventricular ejection fraction = 52 %

我需要提取左心室射血分数的数值

输出应为

ID Lowerbound   Upperbound
1   60            65
2   55            60
3   40            50
4   52            NULL

我已尝试过以下sql语法搜索字符 - 它因ID 3和4而失败（对于ID 4，它给我50但应该是52）

SELECT SUBSTRING(textfield,CHARINDEX('-', 
textfield)-1,10),dbo.udf_GetNumeric(RIGHT(left(textfield,CHARINDEX('- 
',textfield)-1 ),10))AS Lower_bound,left(textfield, CHARINDEX('-', 
textfield) )


,dbo.udf_GetNumeric(SUBSTRING(textfield,CHARINDEX('-', textfield)+1,5)) 
    AS Upper_bound

提前谢谢

Answer 1

从技术上讲，这适用于您的Upperbound问题，至少在您的示例数据的范围内：

CASE WHEN RIGHT(TextField, 4) LIKE '%-[1234567890]' + '%' 
    THEN Left(Right(TextField,3),2) 
WHEN RIGHT(TextField, 6) LIKE '%to [1234567890]%'
    THEN Left(Right(TextField,3),2) 
WHEN RIGHT(TextField, 6) LIKE '%= [1234567890]%'
    THEN Left(Right(TextField,4),2) 
END AS UpperBound

但是，你的示例数据非常糟糕，m

Answer 2

请尝试使用此代码，如果这适用于您的整个数据集，请告知我们。这里的假设是：

1）你的数字总是在“左心室射血分数”字样之后出现。

2）左心室射血分数＆＃39;之后没有其他数字。除了定义下限和上限的值

3）下限和上限始终从左到右（从小到大）

4）起始表的名称是#temp

{{1}}

Answer 3

这会使用几个CROSS APPLYs来越来越近。一些替换使得字符串可比，后端的第一个空白是边框。其余的很容易。

DECLARE @tbl TABLE(ID INT, Textfield VARCHAR(500));
INSERT INTO @tbl VALUES
 (1,'estimated left ventricular ejection fraction 60-65%')
,(2,'estimated left ventricular ejection fraction is 55-60%')
,(3,'Left ventricular ejection fraction is approximately 40 to 50%')
,(4,'Fractional Short 50 %( 28-48) LV mass 83 gm (<220) systolic function  left ventricular ejection fraction = 52 % ');

SELECT ID
      ,Rev
      ,substr
      ,CASE WHEN hyph>0 THEN LEFT(substr,hyph-1) ELSE substr END AS LowerBound
      ,CASE WHEN hyph>0 THEN SUBSTRING(substr,hyph+1,10) ELSE NULL END AS UpperBound
FROM @tbl t
CROSS APPLY(SELECT REVERSE(RTRIM(REPLACE(REPLACE(t.Textfield,' to ','-'),' %','%')))) AS A(Rev)
CROSS APPLY(SELECT PATINDEX('% [^1-9]%',A.Rev)) AS B(pos)
CROSS APPLY(SELECT LTRIM(RTRIM(REPLACE(REVERSE(LEFT(A.Rev,B.pos)),'%','')))) AS C(substr)
CROSS APPLY(SELECT CHARINDEX('-',C.substr)) AS D(hyph);

从基于特定子字符串短语获取数值

3 个答案: