|
Question
|
Answer
|
| I have a changing number of files with similar names and the same data layout that I want to be able to read into a single DATA step. Is there an easy way to read them all without necessarily knowing all their names and how many exist?
|
In SAS version 6.12 and above for Windows and UNIX you can use the wildcard character '*' in the file specification for FILENAME or INFILE, so if all you files begin with the same text the code will read them all, e.g.:
FILENAME indata 'file*.dat';
DATA sasdata1;
INFILE indata;
INPUT column1 column2;
/* more code ... */
RUN;
DATA sasdata2;
INFILE 'data*.txt';
INPUT column1 column2;
/* more code ... */
RUN;
|
| I know which records I want to extract, but how do I split a flat file into several separate files using SAS Software?
|
You can now use the special SAS variables _N_ (input record counter) and _INFILE_ (input record text buffer) to select and copy, respectively, the input records without any changes, e.g.:
FILENAME outdata1 'fileout1.dat';
FILENAME outdata2 'fileout2.dat';
DATA _NULL_;
INFILE 'filein.dat';
INPUT;
IF 23 <= _N_ < 45 THEN DO;
FILE outdata1;
PUT _INFILE_;
END;
ELSE IF 45 <= _N_ < 100 THEN DO;
FILE outdata2;
PUT _INFILE_;
END;
RUN;
|
| Can I read multiple flat files into a single DATA step?
|
Yes, try the following sample code which reads the 3 flat files sequentially:
FILENAME indata
('indata1.dat', 'indata2.dat', 'indata3.dat');
DATA sasdata;
INFILE indata;
INPUT .........;
.......;
RUN;
|
| How do you find a specific string in a long input record, then read the data immediately afterwards?
|
Try the following sample code which looks for the first occurrence of the strings 'MONDAY' and 'TUESDAY', then reads the number following them in 2 different ways:
FILENAME indata 'indata1.dat';
DATA sasdata;
INFILE indata;
dayname='MONDAY';
INPUT @dayname number1 8.
@'TUESDAY' number2 8.;
.......;
RUN;
|
