Exporting a large CSV cuts off variable names

[afrerichs]

When I try to export a dataset with 1900 variables as a CSV, the output only contains a certain number of variable names and then the rest are just blanks. 

I tried this code instead of proc export: 

proc transpose data=final(obs=0) out=names ; run;
data _null_;
file 'myfile.csv' dsd ;
set names end=eof;
put _name_ @;
if eof then put;
run;
data _null_;
set final ;
file 'myfile.csv' dsd mod ;
put (_all_) (+0);
run; 

But it STILL cut off the variable names. How can I fix this? 

[Tom · Accepted solution]

Tell SAS to use a longer line length. 

file 'myfile.csv' dsd lrecl=1000000;

You must have some really strange variable names.  The default record length is 32,767 which for 1,900 means your average variable name must be over 16 characters long!  

1192  data _null_;
1193    x=(32767-1899)/1900;
1194    put x=;
1195  run;

x=16.246315789

Good luck trying to write any code to use such a dataset.

[afrerichs]

Thank you! I was able to get it to work. Yes, unfortunately this dataset has crazy variable names. 

My next question in case you know the answer, I want the output csv to be named "OUTPUT_File.csv", but it is outputting it as "Output_File.csv". Do you know if there is a way around this? 

[Tom]

Since the code you posted is going to make a file named myfile.csv you need show what code you actually are running.

If the "Output" part is coming from a macro variable of some sort perhaps you just need to add a call to the %UPCASE() macro function?

---

@afrerichs wrote:

Thank you! I was able to get it to work. Yes, unfortunately this dataset has crazy variable names. 

My next question in case you know the answer, I want the output csv to be named "OUTPUT_File.csv", but it is outputting it as "Output_File.csv". Do you know if there is a way around this? 

---