Importing raw data

[bat069]

Hi all. I am new to SAS and am currently struggling with importing data. I am trying to import a raw data file that is set out in column format (ie no delimiters). There are 49 variables and approximately 1 million observations.

Here is a copy of my code (with most variables not shown):

data test;

     infile 'C:\datafile';

     input episode 1-12

               umrn $ 13-21

               gender 22-26

               ..........

               discharge_diag $ 252-501 (this is variable 29)

               pro $ 502-521

               .........

               sla06 $ 1186-1194; (variable 49)

run;

The problem I am having starts at variable number 29 which is 250 characters long, and in the vast majority of cases, is empty. When I import, for some reason when it gets to variable 29 it drops to the next observation and every variable after that (30-49) also starts on a new observation. I tried 'missover' at the end of the infile line but this just caused variable 29 - 49 to be empty.

Any help would be greatly appreciated.

Brian

Perth Australia

[Kurt_Bremser]

Use the "truncover" option when you have variable length records. This will cause the last variable(s) to be filled with what's available.

[bat069]

Hi Kurt,

Truncover has the same effect as missover...all the data from variable 29-49 just disappears.

Any other ideas?

Brian

[Kurt_Bremser]

a) could you please post an example of the input data?

b) could it be that you have linefeeds or CR/LF in the text data (the $250 field)? That would cause a new (short) line, and since you read with position, every variable whose pointer lies beyond the EOL is set to missing