求助infile的dlm选项

想请问大家一个问题：

我想利用如下代码完成某种宏，

1. data a;
   
2. infile datalines dlm='\t';
   
3. input a $ b $ c $ d $ e $ f $ @@;
   
4. datalines;
   
5. STUDYID(\t)Study Identifier(\t)Char(\t)15(\t) (\t)EG.STUDYID
   
6. BRTHDTC(\t)Date/Time of Birth(\t)Char(\t)19(\t)data9.(\t)ADSL.BRTHDTC
   
7. AGEGR1(\t)Age Group(\t)Char(\t)15(\t)"<40,
   
8. >=40 and < 60,
   
9. >= 60"(\t)(\t)ADSL.AGEGR1
   
10. AGEGR1N(\t)Age Group (N)(\t)Num(\t)8(\t)(\t)ADSL.AGEGR1N
    
11. ;
    
12. run;

复制代码

出来的数据集应该是这样的：

a	b	Char	d	e	f
STUDYID	Study Identifier	Char	15		EG.STUDYID
BRTHDTC	Date/Time  of Birth	Char	19		ADSL.BRTHDTC
AGEGR1	Age  Group	Char	15	<40,     >=40 and < 60,     >= 60	ADSL.AGEGR1
AGEGR1N	Age  Group (N)	Num	8		ADSL.AGEGR1N

注：<40,后面是个换行符。
注：datalines中的(\t)代表从excel直接复制过来的转义符。

中间的datalines为每次手动黏贴进去的东西。
现在我想把这些数据读入，其中用来分隔的是excel的分隔符：\t，就是按照\t的转义符一口气读下去。

为什么不直接用import，我有如下苦衷：

因为公司用的是unix sas，故中间有些难弄得东西
1. 据悉proc import不能读入xlsx，这是死穴。
2. 就算能读xlsx，还有一大问题：文件名中可能存在空格，例如：‘/user1/abc/a b  c.xlsx’

所以只好出此下策，加入一个手动环节，然后想用中间的过程步来自动识别读入的数据。

还望给与回复，非常感谢。
祝阅者新年快乐，马上有房有车有房车，还没毕业的就马上有SCI~~

更新：
感觉貌似不行，。。。因为datalines会把制表符直接转换成空格，例如：
input a $1. @@;

  if a = '09'x then input '|' @@;
  else  input a $1. @@;
这样都是没效果的。。

那么，能否通过import的 ‘/user1/abc/a b  c.xlsx’ 解决呢。。。

(1) you cannot import XLSX files directly because MS XLSX files are XML based. XML is a nested/hierarchical structure that is very hard for importing (which is designed for tabular structures).
(2) if you save one XLSX into a tab-delimited flat text file, then you have to save it either in a shared location or move it into UNIX in binary mode (the tabs will become blanks in ASCII mode, thus mess up your formatting). An INFILE statement with DLM='09'x and TERMSTR=CRLF should work.

回去试试。

个人感觉有点像这个意思：

infile datalines dlm='prxparse(/\t/)';

如果能有办法把datalines里的内容直接用正则替换也是个办法，但是对于datalines。。。

This code did not completely reach your request. newline should be deleted first.

data a;
infile datalines dlm=','  truncover dsd ;
input @;
_infile_=tranwrd(_infile_,'(\t)',',');
input a $ b :$18. c $ d $ e: $24. f $15.;
datalines;
STUDYID(\t)Study Identifier(\t)Char(\t)15(\t) (\t)EG.STUDYID
BRTHDTC(\t)Date/Time of Birth(\t)Char(\t)19(\t)data9.(\t)ADSL.BRTHDTC
AGEGR1(\t)Age Group(\t)Char(\t)15(\t)"<40,>=40 and < 60,>= 60"(\t)(\t)ADSL.AGEGR1
AGEGR1N(\t)Age Group (N)(\t)Num(\t)8(\t)(\t)ADSL.AGEGR1N
;
run;

farmman60 发表于 2014-1-13 01:38
This code did not completely reach your request. newline should be deleted first.

data a;

Hi farmman,

Thank you to given me powerful support first.
But you just misunderstood my mind.
'(\t)' is not a delimiter which contain '\t' with a pair of brackets.
It's means an invisible mark between two cell, for example, from EXCEL.
When you copy two cells form EXCEL to notepad, it contain a '\t' between two cells.

上面的方法可以的，将xlsx的数据复制保存到TXT，然后用下面的语句读取TXT文件
data dlminput;
        infile '/folders/myfolders/exampractice/input/input02.txt' dlm = '09'x;
        input color $ r ;
run;