Why proc freq is aligning categorical variable vertically rather than horizontally?

[Ruhi]

Hi , 

I ran proc freq on my data (where my remarknew is a character var with certain length). See table below. So remarknew is newly created var and I used a format statement in datastep to change its format to $19. and verified with proc contents. But why is SAS aligning results this way (SAS9.4). I have to change misspelled categories here, but this alignment is just making it annoying to read words. Any help is appreciated?

Proc contents results;

46	remarknew	Char	19	$19.
47	id	Num	8
48	ptnew	Char	15
49	case	Char	19	$19.
50	trial	Char	19

Thanks

~R

remarknew	Frequency	Percent	Cumulative Frequency	Cumulative Percent
B A E L I N E 1	12	0.15	12	0.15
B A S E L I N E 1	407	4.97	419	5.12
B A S E L I N E 2	420	5.13	839	10.25

[PaigeMiller]

What output destination is this?

[Ruhi]

ods html .

[Reeza]

What's your output destination? What is your current setting for linesize?

[Ruhi]

I have not set the linesize, its default. ods HTML is output destination. 

[Reeza]

Is it happening for all PROC FREQ or just this table?

What happens if you run the following?

proc freq data=sashelp.class;
table name;
run;

If it doesn't happen with that, its related to your data, not the PROC. If it does happen, it's a setting in SAS somewhere most likely.

[Ruhi]

What can I do with my data about this issue? 

[Reeza]

Have you ever modified your templates?

[Ruhi]

No I didn't. And its not happening with the data steps you provided. 

[Reeza · Accepted solution]

Then it's a data issue of sorts. I would first try cleaning the data and removing any hard returns using COMPRESS().

data clean;
set old;

newVar = compress(oldVar, ,'s'); *note the two commas;

run;

proc freq data=clean;
table newVar;
run;

---

@Ruhi wrote:

No I didn't. And its not happening with the data steps you provided. 

---

[Ruhi]

This does not work either, I had already tried it. 

[Reeza]

Show the code you tried.
Did you import the data from Excel?
Another method for debugging is to apply a hex format to your data and show us what that looks like.

[Ruhi]

data full_sta;
set full;
where motion_ = "No Motion";
remarknew=input(upcase(compress(remarks,,'s')), $19.);
case=input(compress(remarknew,' ',"d"), $19.);
format remarknew $19. case $19.;
run;

I imported a csv file. I will try hex FORMAT. 

[ballardw]

Can you share the code you used to create the variable? And the exact proc freq code used to create the output?

Is there a specific reason that variable has a $19. Format assigned? And what was the format prior to you changing it? You don't show any values longer than what should be 9 or 10 characters so I'm not sure why the $19 was considered desirable.

[Ruhi]

I added that format to data, based on how SAS read it from file. So remarks was the original column and I added compress and format to create the new variable remarknew as follows. 

data full_sta;
set full;
	where motion_ = "No Motion";
	remarknew=input(upcase(compress(remarks,,'s')), $19.);
	case=input(compress(remarknew,' ',"d"), $19.);
	format remarknew $19. case $19.;
run;

proc freq data=full_sta;
table remarknew;
run;

Variables in Creation Order
#	Variable	Type	Len	Format	Informat
1	pt	Num	8	NLNUM12.	NLNUM32.
2	remarks	Char	19	$19.	$19.