Missing Observations after extracting character values

[michokwu]

Hello experts,

I have survey results like the table below. For ease of analysis,  I extracted the 1st character and converted to numeric cause of the missing values, the first time I ran the code the complete table was created, the second  time there are lots of observations missing, cant seem to find out why, the log does not have an error statement.

Thanks;

/*Sample code*/

data sample;
set have;
new_var=substr(old_var,1,1);
run;

If device=  ' ' then A001=0;
else If device= 'b' then A001=1;
else If device='c' then A001=2;

label A001="device";

run;

Device	active social media	Age	Gender	satisfied with CSR	preferred time to call	Location
	b=No	22	a=1		e= Anytime	b=instate
b=iphone	a=Yes	28	b=2	a=Yes	a=Morning	a=outstate
b=iphone		35		a=Yes	e= Anytime	a=outstate
b=iphone	a=Yes	35	b=2	a=Yes	e= Anytime	b=instate
a=samsung	a=Yes	27		a=Yes	e= Anytime	b=instate
a=samsung	a=Yes	32		a=Yes	a=Morning	b=instate
		37	b=2			b=instate
b=iphone	b=No	28	a=1	a=Yes		b=instate
	b=No	25				a=outstate
	b=No		a=1		a=Morning	a=outstate
		30yrs				b=instate
c=a & b				a=Yes	e= Anytime
	b=No
		23	b=2		c=Afternoon	b=instate
		30	b=2			b=instate
a=samsung	b=No			a=Yes	e= Anytime

I 

[Kurt_Bremser · Accepted solution]

When you run the same code on the same data, you get the same result. When the result changes, you have either changed the code, or used different input data.

Your code as posted is incomplete (everything after the first run; is invalid).

Please use the "little running man" for posting code, and post example data for dataset "have" in a data step with datalines.

[michokwu]

Thanks. Yes, I noticed this after editing it a little. It's the same data, I renamed some variables but followed the same coding convention that I used before. 

Apologies! this is what my code actually looks like. 

/*Sample code*/
data sample;
set have;
new_var=substr(old_var,1,1);
run;

 
data sample1;
set sample;
If device=  ' ' then A001=0;
else If device= 'a' then A001=1;
else If device= 'b' then A001=2;
else If device='c' then A001=3;
label A001="device";

keep A001;
run;

[ed_sas_member]

Hi @michokwu 

It is seems that you should replace device by new_var:

data sample;
	set have;
	new_var=substr(old_var, 1, 1);
run;

data sample1;
	set sample;
	if new_var=' ' then A001=0;
	else if new_var='a' then A001=1;
	else if new_var='b' then A001=2;
	else if new_var='c' then A001=3;
	label A001="device";
	keep A001;
run;

Best,

[michokwu]

I changed the names of new variables back to the original names.

Thank you

[michokwu]

It was the code. An 'else if' statement was not specified correctly