DATA Step, Macro, Functions and more

How do I Identify Palindromes?

Reply
New Contributor
Posts: 2

How do I Identify Palindromes?

[ Edited ]

I've been attempting a programming exercise that I found online, but there are no solutions posted. The prompt is as follows: Given any input dataset, identify any palindromes that exist within the set. 

 

There are two types of palindromes: case sensitive, and case insensitive. I am a recent SAS programmer and am looking to hone my skills through new exercises! 

 

 

Super User
Posts: 6,543

Re: How do I Identify Palindromes?

[ Edited ]
Posted in reply to austinrousevelt

Avoid macros entirely for this problem.  Macros just automate a working process.  They are of little use unless you can program the result without using macros.

 

Take a look at three functions:  REVERSE, STRIP, and UPCASE.

 

Make that four functions.  You'll also need COMPRESS.  (Once you have COMPRESS, you may not need STRIP anymore.)

 

Hint:  this will be a very short program, once you apply the right tools in the right way.

 

Able was I ere I saw Elba

PROC Star
Posts: 1,344

Re: How do I Identify Palindromes?

Posted in reply to austinrousevelt
data pal;
     input word $25.;
     datalines;
Eye
nicola
Pop
roger
debra
sunil
veena
mirna
Noon
Level
Radar
Kayak
Rotator
john
malayalam
mark
steve
keith
naveen
jacob
Brett
Poornima
Badrinarayanan
Nepun
Parthasarathy
;
 
/*Below is the logic to check for Palindromes*/
data want;
     set pal;
     length result $30;
 
     if mod(length(strip(word)),2)=0 then
           do;
                do i=1 to divide(length(strip(word)),2);
                     if substr(lowcase(word),i,1) ne substr(lowcase(reverse(strip(word))),i,1) then
                           do;
                                result=' Not palindrome';
                                output;
                                go to next;
                           end;
                     else result='palindrome';
 
                     if i=divide(length(strip(word)),2) and result='palindrome' then
                           output;
                end;
           end;
     else if mod(length(strip(word)),2)=1 then
           do;
                do i=1 to median(1,length(strip(word)))-1;
                     if substr(lowcase(word),i,1) ne substr(lowcase(reverse(strip(word))),i,1) then
                           do;
                                result='Not palindrome';
                                output;
                                go to next;
                           end;
                     else result='palindrome';
 
                     if i=median(1,length(strip(word)))-1 and result='palindrome' then
                           output;
                end;
           end;
 
next:
run;
Super Contributor
Posts: 500

Re: How do I Identify Palindromes?

Posted in reply to austinrousevelt

The specification is extreme sparse. So first thing to do is creating an input-dataset and a second dataset with the expected solution. This allows you to use proc compare to verify that the generated solution matches the expected result. As soon as you have those two datasets we can think about how to solve this.

PROC Star
Posts: 229

Re: How do I Identify Palindromes?

Posted in reply to austinrousevelt

I would use PROC FCMP to write a function, e.g.:

proc fcmp outlib=work.funcs.str;
  function palindrome(str $,options $);
    str=left(str);
    options=upcase(options);
    if index(options,'I') then
      str=upcase(str);
    l=length(str);
    if l=1 then
      return(1);
    half=floor(l/2);
    if substr(str,1,half)=reverse(substr(str,l-half+1,half)) then
      return(1);
    return(0);
  endsub;
run;
options cmplib=work.funcs;
data test;
  input str $;
  p=palindrome(str,'');
  pi=palindrome(str,'i');
cards;
a
aga
aggA
aggb
;run;
Super User
Posts: 10,621

Re: How do I Identify Palindromes?

Posted in reply to austinrousevelt

It is more like Computer Science student's question. Do you major CS ?

And I highly suggest to use REVERSE() proposed by @Astounding

 

data _null_;
x="Able was I ere I saw Elba Hint:  this will be a very short program, once you apply the right tools in the right way";
do i=1 to length(x);
 do j=i+1 to length(x);
   temp=substr(x,i,j-i+1); 
   if temp=strip(reverse(temp)) and length(temp)>4 
   /* 4 is the Palindromes's least length, you could change it.*/
   then put 'Found: ' temp=;
 end;
end;
run;
Super User
Posts: 10,621

Re: How do I Identify Palindromes?

Posted in reply to austinrousevelt

If you want case insensitive, put UPCASE() around it.

 

   if upcase(temp)=upcase(strip(reverse(temp))) and length(temp)>4 
Community Manager
Posts: 3,362

Re: How do I Identify Palindromes?

This is fun. I've augmented @Ksharp's approach to add case sensitive/insensitive checks, and also eliminate word spaces/punctuation from affecting the result.  After all...

palindrome is a wordphrase, number or sequence of words that reads the same backwards as forwards. Punctuation and spaces between the words or lettering is allowed.

 

data _null_;
  x="Able was I ere I saw Elba Hint:  this will be a very short program, once you apply the right tools in the right way";
  /* 4 is the Palindromes's least length, you could change it.*/
  min = 4;
  do i=1 to length(x);
    do j=i+1 to length(x);
      original = substr(x,i,j-i+1);
      temp=compress(original,,'ka');
      temp_ci=upcase(compress(original,,'ka'));
      if length(original) >= min then
        do;
          if temp=strip(reverse(temp)) then
            put 'Case sensitive: ' original=;
          if temp_ci = strip(reverse(temp_ci)) then
            put 'Case insensitive: ' original=;
        end;
    end;
  end;
run;

I'll bet that more improvements are possible.

Ask a Question
Discussion stats
  • 7 replies
  • 335 views
  • 5 likes
  • 7 in conversation