[Okta Webinar] Learn how to a build a cloud-first strategyRegister Now

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 854
  • Last Modified:

Algorithm of files comparison

I am now writing a java program which compares 10 files for producing the difference, so that the user can change the difference easily. This is similar to a software WinMerge which can compare 2 files only. I want to ask whether there is any similar algorithm that some people have met and found some in web.
Moreover, this program needs the comparison of different language, it seems that different language should have different algorithm.?At this time, I need to support the English and Chinese (in unicode). However, I want to know whether I need to implement in 2 algorithm for these 2 languages. How about the other?

Moreover, if there is Chinese and English at the same time in the file, how to deal with the comparison?


Thanks!
0
fungho
Asked:
fungho
1 Solution
 
rajesh_balaCommented:
listening
0
 
funghoAuthor Commented:
listening?
0
 
dnoelppCommented:
rajesh doesn't know the answer, but is interested in the answer...

Rajesh, this is not neccessary, just press on the subscribe button to receive e-mail notifications.
0
What does it mean to be "Always On"?

Is your cloud always on? With an Always On cloud you won't have to worry about downtime for maintenance or software application code updates, ensuring that your bottom line isn't affected.

 
dnoelppCommented:
And, fungho, as for the different languages in Unicode, there's not really a problem. Each "letter" or better said, a "symbol" is represented by a char value. Just to the comparison on char-values. This way, the Latin letter "a" will be recognized being different from a Chinese symbol.

Say, you have two files:

--- file one ---
Hello, World!
--- file one ---

and

--- file two ---
Hello, <Chinese Symbol for Chinese>World!
--- file two ---

then the symbol in front of the big Latin letter 'W' (the chinese symbol for "Chinese") is treated as a single "letter". So, this way, the second file just differs in the additional letter <Chinese Symbol for Chinese>.
0
 
omry_yCommented:
1. what exactly are you trying to achive?
2. why different algorithns for different languages?
0
 
funghoAuthor Commented:
It needs different algorithm is that for English, there is space between each word, however, for the Chinese, there is no space. So I need to cut the Chinese word by characters.
0
 
dnoelppCommented:
Now I understand your problem a little better. Do you want to do sort of an overview: number of each word, so say in a text the word "I" is counted 4 times, "a" 5 times, "times" 3 times, etc, do you understand? To achieve this, we can use a sorted hashmap for the words, but that's not the problem. You want to know how to extract words out of the text and consider the differences in punctuation in Chinese and English.

There's the class Character.UnicodeBlock. One of the instances is CJK_UNIFIED_IDEOGRAPHS. I think that this is the instance for all Chinese symbols.

So, to find out whether a character is a Chinese symbol, use this:

Character.UnicodeBlock.of(c) == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS.

When a character is chinese, just dump it in the hashmap for counting.

Then, to find out whether the character is a Latin letter, use this:

Character.UnicodeBlock.of(c) == Character.UnicodeBlock.BASIC_LATIN.

I am not sure, however, whether BASIC_LATIN covers accented letters like a e i ? u c, that's up to you to find out.

When you have a latin letter, it can be a part of a word. How to find out? I suggest to partition the Unicode character set into three parts: 1. Chinese symbols 2. Latin letters 3. None of the first two.

Algorithm:

Read the file character-wise. Start with an empty word (for English). For each character do:

1. Chinese symbol? Dump it into the hash map!
2. Letter? Add it to the word.
3. None of the two? Is the word emtpy, don't do anything. Else dump the word to the hash map and make it empty.

Link:

www.unicode.org
0
 
funghoAuthor Commented:
Thanks! dnoelpp! In fact, I have solved the problem by myself. But I check the character whether it is english or chinese by using this method: ( < "\u00FF")... I will use your method tomorrow. Besides I have a question also, if there is a language which starts from right to left, at the same time, there is number and english embeded in it, so some words needs to be read from left to right? how to deal with this? This does not mind whether you know or not... Thanks!
0
 
dnoelppCommented:
That's a problem for Swing. Some Swing components, say, JLabel have to cope with it. Imagine an editor where you edit arabic and english text at the same time. What happens if you move the cursor from the english part to the arabic part.

Strings or Unicode sequence, however, they follow the principle that the "reading" order is important, not the placing. An example. Two words, "english" and "arabic" are in a string, and "arabic" is printed from right to left. Then the string would be displayed as such:

english cibara

But the string sequence is as follows: 'e' 'n' 'g' 'l' 'i' 's' 'h' ' ' 'a' 'r' 'a' 'b' 'i' 'c'
0
 
funghoAuthor Commented:
Then does the java swing deal with the situation that the arabic comes from right to left? or I need to deal with it?
0
 
dnoelppCommented:
Swing does. Please read this article:
http://java.sun.com/products/jfc/tsc/articles/bidi/index.html

The keyword is "bidi" (short for bi-directional text)

By the way, this could be interesting for you (collating and sorting in other languages than English):
http://java.sun.com/docs/books/tutorial/i18n/text/index.html
0
 
funghoAuthor Commented:
Thanks for your help! I will read it later!
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Tackle projects and never again get stuck behind a technical roadblock.
Join Now