What is the best way to remove multiple occurrences of a character in a string in java

Question

I have a string like foo..txt and I want to convert it to foo.txt The occurrence of '.' maybe more than 2. What is the best way to do this?

edit: "." may not happen only together. Entries may also be lower:

foo.bar.txt = foo bar.txt
foo..bar.foo.txt = foo bar.txt

Answer 1

I believe that you want to replace all periods in the file name part with spaces, but keep the extension, right?

If so, then something like this would be appropriate:

String[] tests = { "foo.bar.txt", // [foo bar.txt] "foo...bar.foo.txt", // [foo bar foo.txt] "........", // [.] "...x...dat", // [x.dat] "foo..txt", // [foo.txt] "mmm....yummy...txt" // [mmm yummy.txt] }; for (String test : tests) { int k = test.lastIndexOf('.'); String s = test.substring(0, k).replaceAll("\\.+", " ").trim() + test.substring(k); System.out.println("[" + s + "]"); } 

Essentially how it works:

• First find lastIndexOf('.') In our line
  • Let's say that this index is k , then we logically divided our line into:
    • substring(0, k) , prefix part
    • substring(k) , part of the suffix (file extension)
• Then we use the regular expression in the prefix part to replaceAll matches \.+ With " "
  • That is, the literal dot \. repeated one or more times +
  • We also trim() this line to remove leading and trailing spaces
• We want a converted prefix combined with the original suffix

Explanation

• The reason the pattern \.+ Instead of .+ Is because the period . is a regular expression metacharacter, but in this case we really mean the literal period, so it should be avoided as \.
• The reason this Java string literal pattern is "\\.+" Is because \ itself is the escape character of a Java string string. For example, the string literal "\t" contains a tab character. Similarly, the string literal "\\" contains a backslash character; it has length() one.

References

• regular-expressions.info/The Dot Matches (Almost) Any Character and Repeat with Star and Plus
• String API : lastIndexOf and trim()
• JLS 3.10.6 Output Sequences for Character and String Literals

Answer 2

With replaceAll() ! Like this:

 string = string.replaceAll("\\.{2,}", ".") 

Note that we had to avoid the period, as it is a special character in regular expressions (and also avoids the backslash for Java). Also pay attention to {2,} , which means "a match if it occurs two or more times."

Answer 3

You made me read the manuals :) I solved a more general problem: how to replace any 2+ identical characters one by one with just one character:

 String str = "assddffffadfdd..o"; System.out.println (str.replaceAll("(.)\\1+", "$1")); 

Conclusion:

 asdfadfd.o 

If you only need a solution for the case of "filename .... ext", then I would prefer something simpler, as in Etaouin's answer, because it probably works faster (but not fact). My solution, simplified for this particular case, is as follows:

 str.replaceAll("(\\.)\\1+", "$1") 

Answer 4

 "file....txt".replaceAll("\\.\\.+",".") 

Regex that matches all occurrences of more than one point and replaces it with one point.

Answer 5

use replaceAll() as follows:

 string = string.replaceAll("\\.+(?=.*\\..*)", " ") 

To paraphrase a regex from left to right:

• "\\.+" Find one or more periods
• "(?=.*\\..*)" after viewing and searching for a period

It handles the test case that you mentioned - for example, cases like:

• Txt.
• test..txt
• .test.txt

are converted as follows:

• Txt.
• test.txt
• test.txt

Answer 6

I suggest String.replaceAll .