count c

[Naveen1111]

Hi, 

i have used the talent dataset. Could someone tell me why there is a difference between x and y results?

data tp;
set sasuser.talent;
blank=strip(Agency);
x=countc(blank," ");
y=countc(strip(Agency)," ");

keep blank x y Agency;

run;

results

Obs Agency blank x y

1 TALENT TALENT TALENT TALENT TALENT TALENT 2 2
2 GEORGE STICKS GEORGE STICKS 8 1
3 UNIVERSAL UNIVERSAL 11 0
4 HORIZONS HORIZONS 12 0
5 SYCAMORE SYCAMORE 12 0
6 GOLDEN GUILD GOLDEN GUILD 9 1
7 GOLDEN GUILD GOLDEN GUILD 9 1
8 ACTORS SQUARE ACTORS SQUARE 8 1
9 LAURELS LAURELS 13 0
10 ARTS TALENT ARTS TALENT 10 1
11 FINDERS FINDERS 13 0
12 ALL STARS ALL STARS 12 1
13 ALL STARS ALL STARS 12 1
14 ALL STARS ALL STARS 12 1
15 UNLIMITED UNLIMITED 11 0
16 STARS AND STRIPES STARS AND STRIPES 5 2
17 SETTINGS SETTINGS 12 0
18 STARS AND STRIPES STARS AND STRIPES 5 2
19 UNIVERSAL LIGHTS UNIVERSAL LIGHTS 5 1
20 MODELS AND MAGIC MODELS AND MAGIC 6 2
21 HORIZONS HORIZONS 12 0
22 HARDY STELSON HARDY STELSON 8 1
23 AKA TALENT AKA TALENT 11 1
24 MODELS AND MAGIC MODELS AND MAGIC 6 2

[PeterClemmensen]

Because you have leading and/or trailing blanks in your input variable. The Strip Function removes them

data test;
   a  = ' abc  ';
   ca = countc(a, ' ');
   cb = countc(strip(a), ' ');
run;

[Naveen1111]

indeed, so on that basis, results should be identical as in blank variable i used strip function, which removed the leading and trailing blanks. But when i used x variable (x=countc(blank," ");)with countc function, still showing spaces, not in y variable case(y=countc(strip(Agency)," ");)

[PeterClemmensen]

I'm not sure what you mean by this?

[ballardw]

---

@Naveen1111 wrote:

indeed, so on that basis, results should be identical as in blank variable i used strip function, which removed the leading and trailing blanks. But when i used x variable (x=countc(blank," ");)with countc function, still showing spaces, not in y variable case(y=countc(strip(Agency)," ");)

---

The variable Blank, when created is the same length as agency. As such it will be padded at the end with blanks.

Example below shows the result of a "stripped" variable sandwiched between : characters. Countc counts them unless explicitly stripped at the time of counting.

data example;
   x='   Some string with lead and trail blanks  ';
   blank=strip(x);
   string= ':'||blank||':';
   y=countc(string,"");
   put string=;
run;

Look in the log to see value of string.