在SPSS或Excel中标准化异质年龄数据
我正在试图使用SPSS / SPSS Syntax / Excel来标准化一列年龄数据（例如到几岁/几个月）。 我的直觉是使用一系列DO IF循环，即：
DO IF CHAR.INDEX(Age, "y")>1... for years DO IF CHAR.INDEX(Age, "m")>1... for months DO IF CHAR.INDEX(Age, "d")>1... for days 并且让程序参考紧接在string之前的数字作为年/月/日的数量，并将其加到一个新的variables中，这个variables可能是日后（最小的单位），然后可以转换为年。
例如，对于一个单元格“3年5个月”：将3 * 365 + 5 * 30.5 = 1248天旧添加到一个新的variables（如“DaysOld”）。
单元格内容的示例（没有任何string的数字假定为年）：
2 5 months 11 days 1.7 13 yr 22 yrs 13 months 10 mo 6/19/2016 3y10m 10m 12y 3.5 years 3 years 11 mos 1 year 10 months 1 year, two months 20 Y 13 y/o 3 years in 2014
下面的语法将解决很多情况，但绝对不是全部（例如“1.7”或“2014年的3年”）。 你需要做更多的工作，但这应该让你开始好起来…
首先，我重新创build您的示例数据来处理：
data list list/age (a30). begin data "2" "5 months" "11 days" "1.7" "13 yr" "22 yrs" "13 Months" "10 mo" "6/19/2016" "3y10m" "10m" "12y" "3.5 years" "3 YEARS" "11 mos" "1 year 10 months" "1 year, two months" "20 Y" "13 y/o" "3 years in 2014" end data.
现在开始工作：
* some necessary definitions. string ageCleaned (a30) chr (a1) nm dmy (a5). compute ageCleaned="". * my first step is to create a "cleaned" age variable (it's possible to manage without this variable but using this is better for debugging and improving the method). * in the `ageCleaned` variable I only keep digits, periods (for decimal point) and the characters "d", "m", "y". do if CHAR.INDEX(lower(age),'ymd',1)>0. loop #chrN=1 to char.length(age). compute chr=lower(char.substr(age,#chrN,1)). if CHAR.INDEX(chr,'0123456789ymd.',1)>0 ageCleaned=concat(rtrim(ageCleaned),chr). end loop. end if. * the following line accounts for the word "days" which in the `ageCleaned` variable has turned into the characters "dy". compute ageCleaned=replace(ageCleaned,"dy","d"). exe. * now I can work through the `ageCleaned` variable, accumulating digits until I meet a character, then assigning the accumulated number to the right variable according to that character ("d", "m" or "y"). compute nm="". loop #chrN=1 to char.length(ageCleaned). compute chr=char.substr(ageCleaned,#chrN,1). do if CHAR.INDEX(chr,'0123456789.',1)>0. compute nm=concat(rtrim(nm),chr). else. if chr="y" y=nm. if chr="m" m=nm. if chr="d" d=nm. compute nm="". end if. end loop. exe. * we now have the numbers in string format, so after turning them into numbers they are ready for use in calculations. alter type dmy (f8.2). compute DaysOld=sum(365*y, 30.5*m, d).