topic Re: input text over varying number of lines no boundaries in SAS Enterprise Guide

input text over varying number of lines no boundaries

mhouse — Mon, 19 Oct 2015 14:59:06 GMT

Hello,

Trying to read in a .txt file that stores information in sections, broken into lines like a paragraph, but the number of lines per section is unknown. There are two blank lines between each section, I really only need the first two lines from each section. Example:

.TXT TITLE LINE
Job_Code1
Job_script information
Job_description_1
_blank
_blank
Job_Code2
Job_script information
Job_description_1
Job_description_2
_blank
_blank
Job_Code3
Job_script information
Job_description_1

This is the code I have so far which reads by line with a line specifier, but the problem is - in this case - job_code2 spans over 4 lines while the rest spans over 3. My code is very simple and I was hoping to keep it neat, maybe using a macro to define the line specifier but I don't have much experience with macros. Any ideas? My code to read by fixed number of lines:

data want;
infile myText firstobs=2 n=5 truncover;
input Job_code $250. #2
Job_script $250. #5;
run;

Re: input text over varying number of lines no boundaries

Astounding — Mon, 19 Oct 2015 15:18:22 GMT

The easiest way is probably to crawl through line by line instead of defining a block of lines. The INPUT logic might read:

input job_code $250.

/ job_script $250.

/ test $10.

/ test $10.;

if test > ' ' then input /;

else input;

drop test;

Of course if there are other variations, you will need to adjust again. But that can be done ... you just need a good grasp of what is in the data.

Good luck.

Re: input text over varying number of lines no boundaries

mhouse — Mon, 19 Oct 2015 22:26:38 GMT

This worked perfectly for my application. Thank you so much! Thank you for the quick help too.