Dot value stored as character

[mikemangini]

Hello,

I have a variable that is formatted $3. Within this variable there is a value '  .' which was causing me some confusion. Initially I thought '.' was the null value for numeric, not character. However it seems this value has 2 spaces in front of it, hence '  .' , which lead me to believe it was a character

Also, when I enter the following tests:

if variable = '  .';

if variable =.;

Both return true.

Please help me to understand what this value is that I have here. What is a practical approach from a data hygiene perspective for cleaning this? How is a character variable identified as being equal to both a character value and a numeric value at the same time?

Best regards,

Mike

[data_null__ · Accepted solution]

They both test true because variable is converted to numeric when tested against missing numeric dot.  See note below. 

As far as "what it is" it is exactly what it appears to be,  2 space and a period.  You can test and set to blank if LEFT(VARIABLE) eq '.' then variable='  ';

data _null_;
   variable = '  .    ';
   y = variable eq '  .';
   z = variable eq .;
   put _all_;
   run;

NOTE: Character values have been converted to numeric values at the places given by: (Line):(Column).
      35:8  
variable=. y=1 z=1 _ERROR_=0 _N_=1

[mikemangini]

Thank you, this clarification is helpful. I didn't realize SAS would automatically convert to numeric like that. I overlooked it in the log. This is valuable information to have.