Replacing all sets of words in string with other sets of words

Question

I am trying to flip every hebrew set of characters inside a string. Lets say I have this string(Instead of hebrew letters, I will be using symbols):

§♀♠♪ this is my message♣♠♦►♣

(You can probably tell which character is in which language). And I want this character set - §♀♠♪ to be replaced with ♪♠♀§.

But, I want message♣♠♦►♣ to be replaced with message♣►♦♠♣, so only the english word inside this will stay unreversed.

How can I do that? (Yes, I know I cant use these symbols in a regular string but this is an example.)

Is there a range in Unicode where the characters you are looking for all exist? — Jason Sperske, Mar 21 '14 at 17:28
you can verify the ASCII codes for all the alphabets from aA to zZ and rectify them from the given string. — Aditya Peshave, Mar 21 '14 at 17:30
First, can you explain how `♣►♦♠♣` is an English word? Also, you seem confused about [bidi](http://en.wikipedia.org/wiki/Bi-directional_text). — Elliott Frisch, Mar 21 '14 at 17:31
@ElliottFrisch OP isn't calling it an English word, he's using that as a placeholder in his example for Hebrew characters. — JWiley, Mar 21 '14 at 17:37
refer this link for the ASCII codes.. http://www.ascii.cl/htmlcodes.htm — Aditya Peshave, Mar 21 '14 at 17:37
Can you post example with Hebrew text with expected output so we could actually do some test before posting answer? — Pshemo, Mar 21 '14 at 17:39
Is the string originally written in reverse Hebrew and you are trying to un-reverse it? `\p{BidiClass:R}` — , Mar 21 '14 at 17:41
@JWiley I think I get what's he asking for now. Elliott (אֱלִיָּהוּ), but that will require people type the hebrew in backwards. — Elliott Frisch, Mar 21 '14 at 17:41
@JasonSperske Actually yes, there is (The letters are אבגדהוזחטיכךלמםנןסעפצץקרשת btw): \u05d0 to \u05d9,\u05da to \u05df, \u05e0 to \u05e2,\u05e4 to \u05e9 and \u05ea. — NonameSL, Mar 21 '14 at 21:17
@ElliottFrisch that is EXCACTLY why I want to reverse the hebrew letters - to create a method that takes the hebrew words/letter combination and FLIP them, only them. — NonameSL, Mar 21 '14 at 21:18
@sln no, it is originally written in unreversed hebrew and I am trying to reverse it.. — NonameSL, Mar 21 '14 at 21:19
So you could combine this answer http://stackoverflow.com/questions/3835917/how-do-i-specify-a-range-of-unicode-characters with some code to make something. — Jason Sperske, Mar 21 '14 at 23:45

Pshemo · Accepted Answer · 2014-03-21T17:51:39.050

1

_{This solution is based on example provided by OP (the one with ♣♠♦►♣) but wasn't tested on real data.}

You should be able to find sequence of two or more Hebrew characters via \p{InHebrew}{2,}.
When you will find them you can use String#reverse method to reverse them.
Last thing is to use appendReplacement and appentTail from Matcher to create new string with updated matched parts.

Here is example which should do what you want

String yourString = ...;//place for your string
Pattern p = Pattern.compile("\\p{InHebrew}{2,}");
Matcher m = p.matcher(yourString);

StringBuffer sb = new StringBuffer();

while(m.find()){
    m.appendReplacement(sb, new StringBuilder(m.group()).reverse().toString());
}
m.appendTail(sb);

String reversedSpecial = sb.toString();
System.out.println(reversedSpecial);

edited Mar 21 '14 at 17:51

answered Mar 21 '14 at 17:45

Pshemo

122,468
25
185
269

@sln Lawyers always told me to use small caps when saying important things and you want people to read them :) – Pshemo Mar 21 '14 at 17:58
1

Hey, hey .. Its supposed to flash for 3 seconds then dissapear. – Mar 21 '14 at 18:03
Thank you so much, this works! But instead of spoon feeding, there is one thing I would like to know and didn't understand... Whats the {InHebrew} part and how did it get the hebrew letters, or if it didn't, then what DID get the heberw letters? – NonameSL Mar 22 '14 at 16:15
If you take a look at [Pattern documentation](http://docs.oracle.com/javase/8/docs/api/java/util/regex/Pattern.html) you will find example with `\p{InGreek}` which will match "A character in the Greek block". Now if you take a look at [block definition](http://docs.oracle.com/javase/8/docs/api/java/util/regex/Pattern.html#ubc) you will see it is `InXXX` where `XXX` part is valid argument of `Characters.UnicodeBlock.forName(...)` and taking closer look at source code of this class will let us know that `Hebrew` is defined there as unicodes from range `0590 - 05FF`. – Pshemo Mar 22 '14 at 16:34
So basically `\p{InHebrew}` is similar to `[\u0590-\u05FF]` character class. – Pshemo Mar 22 '14 at 16:36

score 0 · Answer 2 · answered Mar 21 '14 at 17:34

assume there's an output buffer keeping the final string: when encountering a hebrew character, read it onto a stack, until an english character is found, and then pop out all letter(s) in the stack to the output buffer; english letters are moved to the output buffer directly.

Replacing all sets of words in string with other sets of words

2 Answers2