Trimming trailing whitespace

[Bob Reardon]

Is there a method in Java that only removes trailing whitespace? I need to preserve the format of the data I am diaplaying, but the trailing spaces make the lines too long to be properly displayed on the web page.

I am going through some of the regular expressions sites to see if I can find something that might help. Please let me know if you have solved this problem in the past.

Thanks,
-Bob

[marc weber]

I don't know how efficient this is (creating a new String object every time it chops off a space), but this works...

class TrimEnd{
   public static void main(String[] args){
      String s = "   three prefix spaces and four trailing    ";
 
      //demonstrate "before"...
      System.out.println(s + "<<end");
 
      //remove trailing whitespace...
      while(s.endsWith(" ")){
         s = s.substring(0, s.length()-1);
      }
 
      //demonstrate "after"...
      System.out.println(s + "<<end");
   }
}

(It would be nice to use a regex to find the lastIndexOf anything that's not a space, and then use that value to create a new substring in one call, but I don't see any lastIndexOf methods taking a regex argument.)
[ October 15, 2004: Message edited by: marc weber ]

[Ernest Friedman-Hill]

Can you just use the API method String.trim() (which removes both leading and trailing whitespace?)

[marc weber]

Oh, duh... This is much better. Find the last character that's not a space, then return a substring based on that index.

class Trimmer{
   String trimEnd(String s){
      int i = s.length()-1;
      while(s.charAt(i)==' '){
         i--;
      }
      return s.substring(0, i+1);
   }
 
   //here's a demo of the above method...
   public static void main(String[] args){
      Trimmer t = new Trimmer();
      String s = "   three leading spaces and four trailing    ";      
      t.trimEnd(s);      
      System.out.println(t.trimEnd(s) + "<<ends here");
   }
}

(Yeah, I'm assuming that for some reason you want to keep any leading whitespace, so trim() won't work. Right?)
[ October 15, 2004: Message edited by: marc weber ]

[Stan James]

Marc, I made something very similar that will remove leading and/or trailing instances of any character. So this would be:

trimmed = StringUtil.strip(input, "T", " ");

I borrowed the concept and syntax from the REXX language, which allows any abbreviation of "TRAILING" in the second argument, eg "TR"

[marc weber]

Originally posted by Stan James:
trimmed = StringUtil.strip(input, "T", " ");

...allows any abbreviation of "TRAILING" in the second argument...

Interesting. Is the third argument for "LEADING"? (Maybe I should add this to my library. I really need to get better about saving code snippets.)

[marc weber]

Here's an improved method that can also handle all blanks (" "), empty Strings (""), and null values...

public static String trimEnd(String s){
if(s == null || s.equals("")){
return "";
}
int i = s.length()-1;
while( (s.charAt(i)==' ') && (i > 0) ){
i--;
}
if( (i == 0) && (s.charAt(0) == ' ') ){
return "";
}
return s.substring(0, i+1);
}

[ October 16, 2004: Message edited by: marc weber ]

[Layne Lund]

This is probably a good exercise to just learn how to do it. However, in production level code, I suggest that you just use String.trim() from the standard API instead.

Layne

[marc weber]

Originally posted by Layne Lund:
This is probably a good exercise to just learn how to do it. However, in production level code, I suggest that you just use String.trim() from the standard API instead.

Layne

Yes, but trim() removes both leading and trailing whitespace. What if you wanted to keep one or the other? (I don't know why, but that was the original problem.)

[Stan James]

Just for grins, the REXX syntax second argument is LEADING, TRAILING or BOTH. The interpreter is very friendly, accepting upper or lower case, any truncation like LEAD or TRAIL or BO. L T and B are far and way the most commonly seen. It's easy to support this kind of thing:

if ("LEADING".startsWith(arg.toUpperCase()) ...

I agree with Layne ... for readability and efficiency use trim() where it meets the requirements. If you installed source with your JDK you might read trim() to see how the Sun guys did it. You could do worse than copying them!

[Layne Lund]

Originally posted by marc weber:

Yes, but trim() removes both leading and trailing whitespace. What if you wanted to keep one or the other? (I don't know why, but that was the original problem.)

Good point.

[marc weber]

Actually, when I first considered this problem (eliminating only the trailing whitespace), my initial idea was to count the number of any leading spaces, then use trim(), then put the leading spaces back.

[Ilja Preuss]

Originally posted by Stan James:
If you installed source with your JDK you might read trim() to see how the Sun guys did it. You could do worse than copying them!

Well, I don't know about trim(), but in many cases you actually could do much better. Take a look at GregorianCalendar...

[Warren Dew]

marc weber:

[I]Here's an improved method that can also handle all blanks (" "), empty Strings (""), and null values...

public static String trimEnd(String s){
if(s == null || s.equals("")){
return "";
}
int i = s.length()-1;
while( (s.charAt(i)==' ') && (i > 0) ){
i--;
}
if( (i == 0) && (s.charAt(0) == ' ') ){
return "";
}
return s.substring(0, i+1);
}

[/I]

Any particular reason not to use the Character.isWhitespace() function?

[marc weber]

Originally posted by Warren Dew:
...Any particular reason not to use the Character.isWhitespace() function?

I had forgotten that the Character wrapper has so many unique methods (compared to the numeric wrappers). Thanks for pointing this out!

[Bob Reardon]

Thanks for all of your help in solving my problem. I ran in to this problem while taking billing information that was generated on our mainframe and displaying it on the web. The lines generted for the billing report were too long and needed to have their trailing spaces trimmed. But, I needed to keep the leading spaces to maintaining the formatting. (See example)

example header line:
<pre>
Date Description Total Minimum
Due
...

</pre>

Thanks again for all of you help.
-Bob

[Stefan Wagner]

What's about:

s = s.replaceAll (" *$", ""); 

[Bob Reardon]

Originally posted by Stefan Wagner:
What's about:

s = s.replaceAll (" *$", ""); 

That Works!! Thanks Stefan!!

[Collin VanDyck]

Originally posted by Stefan Wagner:
What's about:

s = s.replaceAll (" *$", ""); 

Also, you could use:

s = s.replaceAll ("\\s*$","");

Where \s (escaped properly in a String as \\s) is the regex shorthand for whitespace, which will include such characters as tabs.