We Keep Coding

Diff comparison word by word and display changes

Could be marked as duplicated, but I haven't found a propper solution yet.
I need to write a function that compares 2 pieces of text word by word, and prints out the text showing added/deleted/changed words. For example:
StringOriginal = "I am Tim and I am 27 years old"
StringEdited = "I am Kim and I am not that old".
Result: I am Tim Kim and I am 27 years not that old.
Most of the diff algorithms I find tend to compare char by char. this works fine, untill you have a 2 different words on the same index, with mutual chars.
"I am Tim" edited to
"I am Kim"
Results into:
I am TKim
instead of
I am Tim Kim.
Any pointers?

Split by space both StringOriginal and StringEdited. Loop thru each word of StringOriginal comparing it to the same word index from Edited. Every unequal word should be put to a temporary variable and concatenate it to the result only when you get equal word again from the loop. Use StringBuilder in creating the result. Hope this helps

Split both strings by space, join the resulting arrays via Union, then back to string like this:
string[] arr1 = str1.Split(' ');
string[] arr2 = str1.Split(' ');
var merged = arr1.Union(arr2).ToArray<string>();
var mergedString = string.Join(' ', merged);

little bit old fashion, but you can try this.
string StringOriginal = "I am Tim and I am 27 years old";
string StringEdited = "I am Kim and I am not that old";
string[] StringOriginalArray = StringOriginal.Split();
string[] StringEditedArray = StringEdited.Split();
string[] newStringArray = new string[StringOriginalArray.Length + StringEditedArray.Length];
int i = 0;
int io = 0;
int ie = 0;
while (i < newStringArray.Length)
{
if (io < StringOriginalArray.Length)
{
newStringArray[i] = StringOriginalArray[io];
io++;
i++;
}
if (ie < StringEditedArray.Length)
{
newStringArray[i] = StringEditedArray[ie];
ie++;
i++;
}
}
string[] finalArray = new string[newStringArray.Length];
int f = 0;
for (int k = 0; k < newStringArray.Length; k=k+2)
{
finalArray[f++] = newStringArray[k];
if (newStringArray[k] != newStringArray[k+1])
{
finalArray[f++] = newStringArray[k+1];
}
}
Console.WriteLine(String.Join(" ", finalArray));
Output:
"I am Tim Kim and I am 27 not years that old"

been looking for an answer myself to this question.
haven't been able to find a good solution.
came up with the following. but it's not perfect.
public static class DiffEngine
{
private static Regex r = new Regex(#"(?<=[\s])", RegexOptions.Compiled);
public static string Process(ref string TextA, ref string TextB)
{
var A = r.Split(TextA);
var B = r.Split(TextB);
var max = Math.Max(A.Count(), B.Count());
var sbDel = new StringBuilder("<del>");
var sbIns = new StringBuilder("<ins>");
var sbOutput = new StringBuilder();
var aCurr = string.Empty;
var bCurr = string.Empty;
var aNext = string.Empty;
var bNext = string.Empty;
for (int i = 0; i < max; i++)
{
aCurr = (i > A.Count() - 1) ? string.Empty : A[i];
bCurr = (i > B.Count() - 1) ? string.Empty : B[i];
aNext = (i > A.Count() - 2) ? string.Empty : A[i + 1];
bNext = (i > B.Count() - 2) ? string.Empty : B[i + 1];
if (aCurr == bCurr)
{
sbOutput.Append(aCurr);
}
else
{
if (aNext != bNext)
{
sbDel.Append(aCurr);
sbIns.Append(bCurr);
}
else
{
sbDel.Append(aCurr);
sbIns.Append(bCurr);
sbOutput
.Append(sbDel.ToString())
.Append("</del>")
.Append(sbIns.ToString())
.Append("</ins>");
sbDel.Clear().Append("<del>");
sbIns.Clear().Append("<ins>");
}
}
}
A = null;
B = null;
sbDel = null;
sbIns = null;
return sbOutput.ToString();
}
}

C# How can I compare two word strings and indicate which parts are different

For example if I have...
string a = "personil";
string b = "personal";
I would like to get...
string c = "person[i]l";
However it is not necessarily a single character. I could be like this too...
string a = "disfuncshunal";
string b = "dysfunctional";
For this case I would want to get...
string c = "d[isfuncshu]nal";
Another example would be... (Notice that the length of both words are different.)
string a = "parralele";
string b = "parallel";
string c = "par[ralele]";
Another example would be...
string a = "ato";
string b = "auto";
string c = "a[]to";
How would I go about doing this?
Edit: The length of the two strings can be different.
Edit: Added additional examples. Credit goes to user Nenad for asking.

I must be very bored today, but I actually made UnitTest that pass all 4 cases (if you did not add some more in the meantime).
Edit: Added 2 edge cases and fix for them.
Edit2: letters that repeat multiple times (and error on those letters)
[Test]
[TestCase("parralele", "parallel", "par[ralele]")]
[TestCase("personil", "personal", "person[i]l")]
[TestCase("disfuncshunal", "dysfunctional", "d[isfuncshu]nal")]
[TestCase("ato", "auto", "a[]to")]
[TestCase("inactioned", "inaction", "inaction[ed]")]
[TestCase("refraction", "fraction", "[re]fraction")]
[TestCase("adiction", "ad[]diction", "ad[]iction")]
public void CompareStringsTest(string attempted, string correct, string expectedResult)
{
int first = -1, last = -1;
string result = null;
int shorterLength = (attempted.Length < correct.Length ? attempted.Length : correct.Length);
// First - [
for (int i = 0; i < shorterLength; i++)
{
if (correct[i] != attempted[i])
{
first = i;
break;
}
}
// Last - ]
var a = correct.Reverse().ToArray();
var b = attempted.Reverse().ToArray();
for (int i = 0; i < shorterLength; i++)
{
if (a[i] != b[i])
{
last = i;
break;
}
}
if (first == -1 && last == -1)
result = attempted;
else
{
var sb = new StringBuilder();
if (first == -1)
first = shorterLength;
if (last == -1)
last = shorterLength;
// If same letter repeats multiple times (ex: addition)
// and error is on that letter, we have to trim trail.
if (first + last > shorterLength)
last = shorterLength - first;
if (first > 0)
sb.Append(attempted.Substring(0, first));
sb.Append("[");
if (last > -1 && last + first < attempted.Length)
sb.Append(attempted.Substring(first, attempted.Length - last - first));
sb.Append("]");
if (last > 0)
sb.Append(attempted.Substring(attempted.Length - last, last));
result = sb.ToString();
}
Assert.AreEqual(expectedResult, result);
}

Have you tried my DiffLib?
With that library, and the following code (running in LINQPad):
void Main()
{
string a = "disfuncshunal";
string b = "dysfunctional";
var diff = new Diff<char>(a, b);
var result = new StringBuilder();
int index1 = 0;
int index2 = 0;
foreach (var part in diff)
{
if (part.Equal)
result.Append(a.Substring(index1, part.Length1));
else
result.Append("[" + a.Substring(index1, part.Length1) + "]");
index1 += part.Length1;
index2 += part.Length2;
}
result.ToString().Dump();
}
You get this output:
d[i]sfunc[shu]nal
To be honest I don't understand what this gives you, as you seem to completely ignore the changed parts in the b string, only dumping the relevant portions of the a string.

Here is a complete and working console application that will work for both examples you gave:
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
namespace ConsoleApplication2
{
class Program
{
static void Main(string[] args)
{
string a = "disfuncshunal";
string b = "dysfunctional";
StringBuilder sb = new StringBuilder();
for (int i = 0; i < a.Length; i++)
{
if (a[i] != b[i])
{
sb.Append("[");
sb.Append(a[i]);
sb.Append("]");
continue;
}
sb.Append(a[i]);
}
var str = sb.ToString();
var startIndex = str.IndexOf("[");
var endIndex = str.LastIndexOf("]");
var start = str.Substring(0, startIndex + 1);
var mid = str.Substring(startIndex + 1, endIndex - 1);
var end = str.Substring(endIndex);
Console.WriteLine(start + mid.Replace("[", "").Replace("]", "") + end);
}
}
}
it will not work if you want to display more than one entire section of the mismatched word.

You did not specify what to do if the strings were of different lengths, but here is a solution to the problem when the strings are of equal length:
private string Compare(string string1, string string2) {
//This only works if the two strings are the same length..
string output = "";
bool mismatch = false;
for (int i = 0; i < string1.Length; i++) {
char c1 = string1[i];
char c2 = string2[i];
if (c1 == c2) {
if (mismatch) {
output += "]" + c1;
mismatch = false;
} else {
output += c1;
}
} else {
if (mismatch) {
output += c1;
} else {
output += "[" + c1;
mismatch = true;
}
}
}
return output;
}

Not really good approach but as an exercise in using LINQ: task seem to be find matching prefix and suffix for 2 strings, return "prefix + [+ middle of first string + suffix.
So you can match prefix (Zip + TakeWhile(a==b)), than repeat the same for suffix by reversing both strings and reversing result.
var first = "disfuncshunal";
var second = "dysfunctional";
// Prefix
var zipped = first.ToCharArray().Zip(second.ToCharArray(), (f,s)=> new {f,s});
var prefix = string.Join("",
zipped.TakeWhile(c => c.f==c.s).Select(c => c.f));
// Suffix
var zippedReverse = first.ToCharArray().Reverse()
.Zip(second.ToCharArray().Reverse(), (f,s)=> new {f,s});
var suffix = string.Join("",
zippedReverse.TakeWhile(c => c.f==c.s).Reverse().Select(c => c.f));
// Cut and combine.
var middle = first.Substring(prefix.Length,
first.Length - prefix.Length - suffix.Length);
var result = prefix + "[" + middle + "]" + suffix;
Much easier and faster approach is to use 2 for loops (from start to end, and from end to start).

String compression for repeated chars

I have the two following strings.
uncompressed "(A(2),I(10),A,A,A,A(3),R,R,R,R,A,A)"
compressed "(A(2),I(10),3A,A(3),4R,2A)"
Ignoring any entries in the format A(n) or I(n) you can see that when we find any consecutive repeated char it is replaced with a single entry at that position and a count.
I know there must be an elegant way to do this but I keep coming up with ugly looking nested loops.
The data in the strings comes from the ISO8211 file format and identify the format to apply to data in the sub fields.
I am sure this could be done with a single line of linq but I am out of ideas (tonight.)

The technique is called Run Length Encoding.
Here's an example using Python:
from itertools import groupby
uncompressed = "(A(2),I(10),A,A,A,A(3),R,R,R,R,A,A)"
counted = [(k, len(list(g))) for k, g in groupby(uncompressed.split(','))]
compressed = ','.join(k if cnt==1 else str(cnt)+k for k, cnt in counted)

Well, not exactly a single line. This will do it:
string str = "(A(2),I(10),A,A,A,A(3),R,R,R,R,A,A)";
string prev = null;
int cnt = 0;
string result =
"(" + String.Join(",",
(str.TrimStart('(').TrimEnd(')') + ",").Split(',').Select(x => {
if (x == prev) {
cnt++;
return null;
} else {
string temp = cnt > 1 ? cnt.ToString() + prev : prev;
prev = x;
cnt = 1;
return temp;
}
}).Where(x => x != null)
) + ")";

Here's a method to do that, using LINQ's GroupBy:
static string RLE(string s) {
s = s.Substring(1, s.Length - 2);
char? l = null;
int i = 0;
return "(" + string.Join(",", s.Split(',').GroupBy(c => {
if(c.Length != 1) {
i++;
return i++;
}
if(c[0] == l) {
return i;
}
l = c[0];
return ++i;
}).Select(x => (x.Count() > 1 ? x.Count().ToString() : string.Empty) + x.First())) + ")";
}

Concatenate neighboring characters of a special character "-"

i am developing an application using c#.net in which i need that if a input entered by user contains the character '-'(hyphen) then i want the immediate neighbors of the hyphen(-) to be concatenated for example if a user enters
A-B-C then i want it to be replaced with ABC
AB-CD then i want it to be replaced like BC
ABC-D-E then i want it to be replaced like CDE
AB-CD-K then i want it to be replaced like BC and DK both separated by keyword and
after getting this i have to prepare my query to database.
i hope i made the problem clear but if need more clarification let me know.
Any help will be appreciated much.
Thanks,
Devjosh

Use:
string[] input = {
"A-B-C",
"AB-CD",
"ABC-D-E",
"AB-CD-K"
};
var regex = new Regex(#"\w(?=-)|(?<=-)\w", RegexOptions.Compiled);
var result = input.Select(s => string.Concat(regex.Matches(s)
.Cast<Match>().Select(m => m.Value)));
foreach (var s in result)
{
Console.WriteLine(s);
}
Output:
ABC
BC
CDE
BCDK

Untested, but this should do the trick, or at the very least lead you in the right direction.
private string Prepare(string input)
{
StringBuilder output = new StringBuilder();
char[] chars = input.ToCharArray();
for (int i = 0; i < chars.Length; i++)
{
if (chars[i] == '-')
{
if (i > 0)
{
output.Append(chars[i - 1]);
}
if (++i < chars.Length)
{
output.Append(chars[i])
}
else
{
break;
}
}
}
return output.ToString();
}
If you want each pair to form a separate object in an array, try the following code:
private string[] Prepare(string input)
{
List<string> output = new List<string>();
char[] chars = input.ToCharArray();
for (int i = 0; i < chars.Length; i++)
{
if (chars[i] == '-')
{
string o = string.Empty;
if (i > 0)
{
o += chars[i - 1];
}
if (++i < chars.Length)
{
o += chars[i]
}
output.Add(o);
}
}
return output.ToArray();
}

Correct me if I am wrong but surely all you need to do is remove the '-'?
like this:
"A-B-C".Replace("-","");

You can even solve this with a one-liner (although a bit ugly):
String.Join(String.Empty, input.Split('-').Select(q => (q.Length == 0 ? String.Empty : (q.Length > 1 ? (q.First() + q.Last()).ToString() : q.First().ToString())))).Substring(((input[0] + input[1]).ToString().Contains('-') ? 0 : 1), input.Length - ((input[0] + input[1]).ToString().Contains('-') ? 0 : 1) - ((input[input.Length - 1] + input[input.Length - 2]).ToString().Contains('-') ? 0 : 1));
first it splits the string to an array on each '-', then it concatenates only the first and the last character of each string (or just the only character if there's only one, and it leaves the empty string if there's nothing there), and then it concatenates the resulting enumerable to a String. Finally we strip the first and the last letter, if they are not in the needed range.
I know, it's ugly, I'm just saying that it's possible..
Probably it's way better to just use a simple
new Regex(#"\w(?=-)|(?<=-)\w", RegexOptions.Compiled)
and then work with that..
EDIT #Kirill Polishchuk was faster.. his solution should work..
EDIT 2
After the Question has been updated, here's a snippet that should do the trick:
string input = "A-B-C";
string s2;
string s3 = "";
string s4 = "";
var splitted = input.Split('-');
foreach(string s in splitted) {
if (s.Length == 0)
s2 = String.Empty;
else
if (s.Length > 1)
s2 = (s.First() + s.Last()).ToString();
else
s2 = s.First().ToString();
s3 += s4 + s2;
s4 = " and ";
}
int beginning;
int end;
if (input.Length > 1)
{
if ((input[0] + input[1]).ToString().Contains('-'))
beginning = 0;
else
beginning = 1;
if ((input[input.Length - 1] + input[input.Length - 2]).ToString().Contains('-'))
end = 0;
else
end = 1;
}
else
{
if ((input[0]).ToString().Contains('-'))
beginning = 0;
else
beginning = 1;
if ((input[input.Length - 1]).ToString().Contains('-'))
end = 0;
else
end = 1;
}
string result = s3.Substring(beginning, s3.Length - beginning - end);
It's not very elegant, but it should work (not tested though..). it works nearly the same as the one-liner above...

Split a string without separator

I have a string variable with value
"abcdefghijklmnop".
Now I want to split the string into string array with, say, three characters (the last array element may contain fewer) in each array element from the right end.
I.e.,
"a"
"bcd"
"efg"
"hij"
"klm"
"nop"
What is the easiest and simplest way to do this?? (Both Visual Basic and C# code is welcome)?

Here's a solution:
var input = "abcdefghijklmnop";
var result = new List<string>();
int incompleteGroupLength = input.Length % 3;
if (incompleteGroupLength > 0)
result.Add(input.Substring(0, incompleteGroupLength));
for (int i = incompleteGroupLength; i < input.Length; i+=3)
{
result.Add(input.Substring(i, 3));
}
It gives the expected output of:
"a"
"bcd"
"efg"
"hij"
"klm"
"nop"

Here's something that works - wrapped up into a string extension function.
namespace System
{
public static class StringExts
{
public static IEnumerable<string> ReverseCut(this string txt, int cutSize)
{
int first = txt.Length % cutSize;
int taken = 0;
string nextResult = new String(txt.Take(first).ToArray());
taken += first;
do
{
if (nextResult.Length > 0)
yield return nextResult;
nextResult = new String(txt.Skip(taken).Take(cutSize).ToArray());
taken += cutSize;
} while (nextResult.Length == cutSize);
}
}
}
Usage:
textBox2.Text = "";
var txt = textBox1.Text;
foreach (string s in txt.ReverseCut(3))
textBox2.Text += s + "\r\n";

Regex time!!
Regex rx = new Regex("^(.{1,2})??(.{3})*$");
var matches = rx.Matches("abcdefgh");
var pieces = matches[0].Groups[1].Captures.OfType<Capture>().Select(p => p.Value).Concat(matches[0].Groups[2].Captures.OfType<Capture>().Select(p => p.Value)).ToArray();
pieces will contain:
"ab"
"cde"
"fgh"
(Please, don't use this code! It is only an example of what can happen when you use a regular expression + LINQ.)

Well... here is yet another way I arrived at:
private string[] splitIntoAry(string str)
{
string[] temp = new string[(int)Math.Ceiling((double)str.Length / 3)];
while (str != string.Empty)
{
temp[(int)Math.Ceiling((double)str.Length / 3) - 1] = str.Substring(str.Length - Math.Min(str.Length, 3));
str = str.Substring(0, str.Length - Math.Min(str.Length, 3));
}
return temp;
}