Convert dataset to txt file

[cloudforest]

I am trying to convert a sas data set to a .txt file.  This seems to be fairly straightforward if I wanted to convert the set into a single text file, but I need each observation to be exported to its own.txt file and have data from two of the variables being the .txt file names.

So if here is my data set:

variable 1       variable 2     variable 3

1                       a                  x

2                       b                  y

3                       c                  z

I want to create 3 text files, 1_a.txt, 2_b.txt and 3_c.txt, containing x, y, z in the text file, respectably.

I've heard this can be done with a sas macro, but I am not familiar with how to write macros and don't know where to start. Any help would be appreciated!

Thanks!

[Tom]

Use the FILEVAR option on the FILE statement and you can do it in a single data step without any macro code.

data _null_;

  set have ;

  _filename=catx('_',variable1,variable2)||'.txt';

  file dummy filevar=_filename ;

  put variable3 ;

run;

[cloudforest]

Thanks for your help!

Using this code, how can I specify the directory I want the txt files to be exported to? When I tried to insert a path to a folder after the file option, I get the error message "A Physical file reference (ie PHYSICAL FILE REFERENCE") or an aggregate file storage reference (ie AGGREGATE(MEMBER)) reference cannot be used with the FILEVAR=option.

[art297 · Accepted solution]

I would guess that it was simply how you defined it.  The following works for me and should for you:

data _null_;

  set sashelp.class ;

  _filename=cats("c:\art\",name,"_",age)||'.txt';

  file dummy filevar=_filename ;

  put variable3 ;

run;

[cloudforest]

It works! Thank you both so much for your help!

[Peter_C]

beware if your data step might write to the same file more than once and the data are not in "filevar" order.

I found the file seemed to be closed when the filevar value changed.

A second write to the file over-wrote the first, if the data are not in "filevar" order.

but here for CLOUDFOREST, it should not be a problem because only one row should be written to each file