Compress array is adding weird characters

[ProcWes]

I run this block before I send data to GCP but for some reason, it is adding weird characters to my strings that aren't there before the compress.

array Chars[*] _character_;
   do i = 1 to dim(Chars);
      Chars[i] = compress(Chars[i],,'kw');
   end;
   drop i;
array Nums[*] _numeric_;
   do i = 1 to dim(Nums);
      Nums[i] = round(Nums[i],0.01);
   end;
   drop i;

These weird characters, the @ etc.  Is there a better way to go about this?  Where are these coming from??

[data_null__]

Are you using DBCS or MBCS.  You might try KCOMPRESS to resolve issue.

---

@ProcWes wrote:

I run this block before I send data to GCP but for some reason, it is adding weird characters to my strings that aren't there before the compress.

array Chars[*] _character_;
   do i = 1 to dim(Chars);
      Chars[i] = compress(Chars[i],,'kw');
   end;
   drop i;
array Nums[*] _numeric_;
   do i = 1 to dim(Nums);
      Nums[i] = round(Nums[i],0.01);
   end;
   drop i;

These weird characters, the @ etc.  Is there a better way to go about this?  Where are these coming from??

---

[Ksharp]

compress(Chars[i],,'kw') would also keep the Chinese,Japanese ..... charaters as a-zA-Z .
Therefore you need to use 'kad' if you only want keep alphabeta（a-zA-Z) and digit(0-9) characters.

Chars[i] = compress(Chars[i],,'kad');

[Patrick]

---

@Ksharp wrote:
compress(Chars[i],,'kw') would also keep the Chinese,Japanese ..... charaters as a-zA-Z .
Therefore you need to use 'kad' if you only want keep alphabeta（a-zA-Z) and digit(0-9) characters.

Chars[i] = compress(Chars[i],,'kad');

---

@Ksharp Wouldn't it need to be Kcompress() ?

[Ksharp]

Patrick,
If OP only need to keep characters (a-zA-Z0-9) , then COMPRESS() and KCOMPRESS() is the same. KCOMPRESS() has the ability to process Chinese,Japanes ....... DBMS characters.

[ProcWes · Accepted solution]

Hi, these replies haven't been solutions to this point.  Compress isn't removing these: ÀõÖ@ because they're showing up as letters and punctuation.  What's troubling to me is that these characters do not exist on the input, even if I convert the string to hex format, they're not there.  The compress is creating them.

Controllable Fraud - Other 436F6E74726F6C6C61626C65204672617564202D204F7468657220202020202020202020202020202020202020202020

[yabwon]

could you share results of this code with us:

data test;
  set <your Data Set> (keep= <one variable that makes problems>);
  test1=<one variable that makes problems>;
  test2 = compress(test1,,'kw');
  a=put(test1,$hex128.);
  b=put(test2,$hex128.);
run;
proc print data = test;
run;

Bart