Related
I am having a bit of a conundrum here... basically I am performing some very very basic file compression steps as follows:
open file and read as string/into a string
parse through the string, and replace repeating patterns with smaller size of text that represents the pattern (ex: aaaaaaaaaaa (11 chars) is replaced with [a#$%11] (8 chars))
save the new, smaller string, into a separate file (can compare sizes)
For some reason, even though the new string in memory is like, 3% smaller than the original string, when I save the string into a file, the file itself is BIGGER than the original file on the file system?? How is that even possible? If someone could explain that to me it would be great!
Here is the code I am using to do this:
void bkg_DoWork(object sender, DoWorkEventArgs e)
{
try
{
string file = File.ReadAllText(this.txt_CompressFilename.Text);
int olength = file.Length;
int nlength = 0;
decimal pct = 0;
string lastchar = "";
int count = 0;
List<RepeatingPattern> SinglePatterns = new List<RepeatingPattern>();
List<RepeatingPattern> DoublePatterns = new List<RepeatingPattern>();
List<RepeatingPattern> TriplePatterns = new List<RepeatingPattern>();
List<RepeatingPattern> QuadruplePatterns = new List<RepeatingPattern>();
UpdateProgress("Read file contents", 0, 1, 6);
UpdateProgress("Finding single character replacements.", pct, 1, 6);
//single character replaces.
for (int i = 0; i < olength; i++)
{
if (file[i].ToString() == lastchar)
count += 1;
else
{
//create a pattern, if the count is more than what a pattern's compressed pattern looks like to save space... 8 chars
//[a#$%#]
if (count > 7)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!SinglePatterns.Contains(ptn))
SinglePatterns.Add(ptn);
}
count = 0;
lastchar = file[i].ToString();
}
}
//handle possible trailing pattern
if (count > 7)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!SinglePatterns.Contains(ptn))
SinglePatterns.Add(ptn);
}
if (SinglePatterns.Count > 0)
for (int i = 0; i < SinglePatterns.Count; i++)
file = file.Replace(SinglePatterns[i].ToString(), SinglePatterns[i].ToReplaceString());
nlength = file.Length;
pct = (decimal)(((double)(olength - nlength) / olength) * 100);
UpdateProgress("Found and replaced " + SinglePatterns.Count, pct, 2, 6);
UpdateProgress("Finding double character replacements.", pct, 2, 6);
lastchar = "";
count = 0;
//double character replaces.
for (int i = 0; i + 1 < file.Length; i = i + 2)
{
if ("" + file[i] + "" + file[i + 1] == lastchar)
count += 1;
else
{
//create a pattern, if the count is more than what a pattern's compressed pattern looks like to save space... 8 chars
//[aa#$%#]
if (count > 8)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!DoublePatterns.Contains(ptn))
DoublePatterns.Add(ptn);
}
count = 0;
lastchar = "" + file[i] + "" + file[i + 1];
}
}
//handle possible trailing pattern
if (count > 8)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!DoublePatterns.Contains(ptn))
DoublePatterns.Add(ptn);
}
if (DoublePatterns.Count > 0)
for (int i = 0; i < DoublePatterns.Count; i++)
file = file.Replace(DoublePatterns[i].ToString(), DoublePatterns[i].ToReplaceString());
nlength = file.Length;
pct = (decimal)(((double)(olength - nlength) / olength) * 100);
UpdateProgress("Found and replaced " + DoublePatterns.Count, pct, 3, 6);
UpdateProgress("Finding triple character replacements.", pct, 3, 6);
lastchar = "";
count = 0;
//triple character replaces.
for (int i = 0; i + 2 < file.Length; i = i + 3)
{
if ("" + file[i] + "" + file[i + 1] + "" + file[i + 2] == lastchar)
count += 1;
else
{
//create a pattern, if the count is more than what a pattern's compressed pattern looks like to save space... 8 chars
//[aaa#$%#]
if (count > 9)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!TriplePatterns.Contains(ptn))
TriplePatterns.Add(ptn);
}
count = 0;
lastchar = "" + file[i] + "" + file[i + 1] + "" + file[i + 2];
}
}
//handle possible trailing pattern
if (count > 9)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!TriplePatterns.Contains(ptn))
TriplePatterns.Add(ptn);
}
if (TriplePatterns.Count > 0)
for (int i = 0; i < TriplePatterns.Count; i++)
file = file.Replace(TriplePatterns[i].ToString(), TriplePatterns[i].ToReplaceString());
nlength = file.Length;
pct = (decimal)(((double)(olength - nlength) / olength) * 100);
UpdateProgress("Found and replaced " + TriplePatterns.Count, pct, 4, 6);
UpdateProgress("Finding quadruple character replacements.", pct, 4, 6);
lastchar = "";
count = 0;
//triple character replaces.
for (int i = 0; i + 3 < file.Length; i = i + 4)
{
if ("" + file[i] + "" + file[i + 1] + "" + file[i + 2] + "" + file[i + 3] == lastchar)
count += 1;
else
{
//create a pattern, if the count is more than what a pattern's compressed pattern looks like to save space... 8 chars
//[aaaa#$%#]
if (count > 10)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!QuadruplePatterns.Contains(ptn))
QuadruplePatterns.Add(ptn);
}
count = 0;
lastchar = "" + file[i] + "" + file[i + 1] + "" + file[i + 2] + "" + file[i + 3];
}
}
//Handle possible trailing pattern
if (count > 10)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!QuadruplePatterns.Contains(ptn))
QuadruplePatterns.Add(ptn);
}
if (QuadruplePatterns.Count > 0)
for (int i = 0; i < QuadruplePatterns.Count; i++)
file = file.Replace(QuadruplePatterns[i].ToString(), QuadruplePatterns[i].ToReplaceString());
nlength = file.Length;
pct = (decimal)(((double)(olength - nlength) / olength) * 100);
UpdateProgress("Found and replaced " + QuadruplePatterns.Count, pct, 5, 6);
UpdateProgress("Saving new .cmp file...", pct, 5, 6);
string newpath = this.txt_FolderName.Text + "\\" + Path.GetFileName(this.txt_CompressFilename.Text);
newpath = newpath.Substring(0, newpath.LastIndexOf("."));
newpath = newpath + ".cmp";
File.WriteAllText(newpath, file);
stopwatch.Stop();
UpdateProgress("Compression completed! Time to compress file: " + string.Format("{0}", stopwatch.Elapsed), pct, 6, 6);
string report = "Compression report\n\n";
FileInfo inf = new FileInfo(this.txt_CompressFilename.Text);
FileInfo infNew = new FileInfo(newpath);
report += "Single character replacements made: " + SinglePatterns.Count + "\n\n";
report += "Double character replacements made: " + DoublePatterns.Count + "\n\n";
report += "Triple character replacements made: " + TriplePatterns.Count + "\n\n";
report += "Quadruple character replacements made: " + QuadruplePatterns.Count + "\n\n";
report += "Total compression ration achieved in string: " + pct + "% \n\n";
report += "Old file size: " + inf.Length + "\nNew file size: " + infNew.Length + " in bytes.";
report += "Total time to achieve compression: " + string.Format("{0}", stopwatch.Elapsed);
e.Result = report;
}
catch (Exception ex)
{
e.Result = ex;
}
}
Here is the code for the RepeatingPattern class...
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
namespace Compressor
{
public class RepeatingPattern : IEquatable<RepeatingPattern>
{
public string RepeatingChar { get; set; }
public int Count { get; set; }
public RepeatingPattern()
{
this.RepeatingChar = "";
this.Count = -1;
}
public RepeatingPattern(string rchar, int count)
{
this.RepeatingChar = rchar;
this.Count = count;
}
public RepeatingPattern(string FromReplaceString)
{
FromReplaceString = FromReplaceString.Replace("[", "").Replace("]", "");
List<string> parts = FromReplaceString.Split(new string[] { "#$%" }, StringSplitOptions.None).ToList();
if (parts.Count != 2)
throw new ArgumentException("Invalid argument count. Must be in this format: [a#$%N]");
try
{
this.RepeatingChar = parts[0];
this.Count = int.Parse(parts[1]);
}
catch (Exception ex)
{
throw new ArgumentException("Unable to cast the argument and create an object from it. Error: " + ex.Message);
}
}
public override bool Equals(object obj)
{
RepeatingPattern tmp = obj as RepeatingPattern;
if (tmp != null)
return base.Equals(tmp);
else
throw new Exception("Invalid comparison type. Both objects must be of type RepeatingPattern");
}
public bool Equals(RepeatingPattern tmp)
{
return this.RepeatingChar == tmp.RepeatingChar && this.Count == tmp.Count;
}
public override int GetHashCode()
{
return this.RepeatingChar.GetHashCode() ^ this.Count.GetHashCode();
}
public override string ToString()
{
string retval = "";
for (int i = 0; i < this.Count; i++)
retval += this.RepeatingChar;
return retval;
}
public string ToReplaceString()
{
return "[" + this.RepeatingChar + "#$%" + this.Count + "]";
}
}
}
Out of curiosity, I have made an attempt at the code. Some differences:
I made a helper function to find runs in the text
I build a new string (using StringBuilder) as I go through the old string instead of replacing in the old string
I think my code is a bit simpler than yours. I have tested with:
Input: "aaaaaaaaaaabbbcdcdcdcdcdcdxxxxxxxxxxxxxxxxxxhello"
Output: "[a#$%11]bbb[cd#$%6][x#$%18]hello"
Here's the code. This is a first draft. Probably lots of improvements to make:
static int FindRun(string s, int start, int length)
{
if (start + length >= s.Length) return 0;
int numRuns = 0;
string pattern = s.Substring(start, length);
for (int i = start + length; i <= s.Length - length; i += length)
{
if (s.Substring(i, length) == pattern) numRuns += 1;
else break;
}
return numRuns;
}
static string EncodeString(string src)
{
StringBuilder sb = new StringBuilder();
for (int i = 0; i < src.Length; i++)
{
string theRun = null;
int numRuns = 0;
// Find runs of lengths 4, 3, 2, 1
for (int j = 4; j >= 1; j--)
{
int runs = FindRun(src, i, j);
if (runs > 1) // Run found!
{
// Save it for later. Want to append the longest run
theRun = src.Substring(i, j);
numRuns = runs;
}
}
// No run? Just append the letter
if (theRun == null)
{
sb.Append(src[i]);
}
else
{
// This is the size of the run
int replacementStringSize = (numRuns * theRun.Length) + (theRun.Length - 1);
// This is the code to use as a replacement
String runCode = String.Format("[{0}#$%{1}]", theRun, numRuns + 1);
// Only append if the code length is smaller than the original run
if (runCode.Length < replacementStringSize)
{
sb.Append(runCode);
}
else
{
// Don't encode. Put original run back
for (int j = 0; j <= numRuns; j++)
{
sb.Append(theRun);
}
}
// Skip over the run
i += replacementStringSize;
}
}
return sb.ToString();
}
The root cause of the much larger output files is because of encoding. ChromeSetup.exe is 1,397,976 bytes. When the file is read in using File.ReadAllText it attempts to detect the string encoding. The string is 1,327,384 characters long in this case. Here's the key though, because of encoding each character isn't necessarily a single byte. For example in UTF-8 each character is 1 to 4 bytes. So then when the result string is written out a single character could become multiple bytes.
For reading/writing executables/binary files you're better off using File.ReadAllBytes()`File.WriteAllBytes()`.
While attempting to run your code I came across several other bugs. Here are the bugs I found.
1) In the double/triple/quad character replaces the for loop bounds should check up to the character that will be used.
//double character replaces.
for (int i = 0; i < file.Length; i = i + 2)
{
if ("" + file[i] + "" + file[i + 1] == lastchar)
This will cause an out of index exception if the file string is an odd number of characters. Add in a + 1 to fix this.
for (int i = 0; i + 1 < file.Length; i = i + 2)
For the triple this will be + 2, for quad + 3.
2) If the string ends with a repeating pattern this isn't handled correctly. In the for loops the pattern count is only checked when a different char is encountered. So if the pattern is at the end of the string it isn't detected. You could handle this by checking the count after the for loop.
if (count > 7)
{
//create and add a pattern to the list if necessary.
RepeatingPattern ptn = new RepeatingPattern(lastchar.ToString(), count);
if (!SinglePatterns.Contains(ptn))
SinglePatterns.Add(ptn);
}
3) count and lastchar should be reset before each for loop. If one for loop ends with count = 17 and the next for loop runs it would add a repeating pattern of count 17, which has already been replaced.
4) As others have mentioned, doing replacements in your input string as you go along has the potential to cause issues.
If you can post your RepeatingPattern code and your input text file we can run down the exact cause of your larger output file.
Edit: Running with your RepeatingPattern code I see another small bug. The pattern "aaaaaaaaaa" becomes "[a#$%9]a". it should be replacing one more character. This could be making your output string slightly longer than expected. To fix this, in the replacement for loops set count to 1 (instead of 0) when a new pattern is started.
I have a List of 9 integers and I want to print all elements from the list as a matrix 3,3. And I have to avoid unnecessary white space on the end of every line.
Is it possible to use String.Join ?
Thanks.
Here's my code:
int[] input = Console.ReadLine().Split().Select(int.Parse).ToArray();
int[][] matrix = new int[input[0]][];
for (int i = 0; i < input[0]; i++)
{
int[] line = Console.ReadLine().Split(' ', StringSplitOptions.RemoveEmptyEntries).Select(int.Parse).ToArray();
matrix[i] = line;
}
List<int> arr = new List<int>(9);
List<int> arr1 = new List<int>(9);
arr = Enumerable.Repeat(0, 9).ToList();
//for (int i = 0; i < 9 ; i++) sum[i%3, i/3] = 0;
for (int row = 0; row < input[0] - 2; row++)
{
for (int col = 0; col < input[1] - 2; col++)
{
arr1.Add(matrix[row][col]);
arr1.Add(matrix[row][col + 1]);
arr1.Add(matrix[row][col + 2]);
arr1.Add(matrix[row + 1][col]);
arr1.Add(matrix[row + 1][col + 1]);
arr1.Add(matrix[row + 1][col + 2]);
arr1.Add(matrix[row + 2][col]);
arr1.Add(matrix[row + 2][col + 1]);
arr1.Add(matrix[row + 2][col + 2]);
if (arr1.Sum() > arr.Sum())
{
arr = arr1.Select(a => a).ToList();
}
arr1.Clear();
}
}
Console.WriteLine($"Sum = {arr.Sum()} ");
// print the list as a matrix
This is how I would print it using String.Join:
List<int> asd = new List<int> {1,2,3,4,5,6,7,8,9};
for (int i = 0; i < asd.Count; i +=3)
{
Console.WriteLine(string.Join(" ",asd.Skip(i).Take(3)));
}
Explanation: Walk in steps of 3. Skip the amount of numbers equal to the stepsize and take 3 to combine a row of the matrix.
You should reconsider the accepted answer because the performance is poorly with many items.
It may be irrelevant for your current count of items but still hear my warning.
I ran the following code snippet:
var sb = new StringBuilder();
for (int i = 0; i < asd.Count; i +=3)
sb.AppendLine(string.Join(" ", asd.Skip(i).Take(3)));
Console.WriteLine(sb.ToString());
used a StringBuilder to remove the time relevant Console.WriteLine(); for every item in the loop.
This approach takes 756,115ms to complete, with 1,000,000 items.
Created the asd list like this:
var asd = Enumerable.Range(0, 1000000).ToList();
Every other answer given so far will perform way better.
The reason why the accepted solution performs this poorly is because of the .Skip() that is getting called inside the loop, it doesn't actually skip and go directly to this Position instead it again and again loops the list till it reaches this point.
My solution would be:
Console.WriteLine(string.Concat(asd.Select((x, i) => (i + 1) % 3 != 0 ? x + " " : x + Environment.NewLine)));
Which executes the same task in 8,610ms
For completness:
Wojtek's solution takes: 7,932ms
Nirmal Subedi' solution takes: 8,088ms
Note:
Changed it so that it uses a StringBuilder to build the string and only output the string once to the console, instead of calling Console.WriteLine() in a loop
Here my complete test routine:
var asd = Enumerable.Range(0, 1000000).ToList();
var sw1 = new Stopwatch();
sw1.Start();
Console.WriteLine(string.Concat(asd.Select((x, i) => (i + 1) % 3 != 0 ? x + " " : x + Environment.NewLine)));
sw1.Stop();
var sw2 = new Stopwatch();
sw2.Start();
var sb1 = new StringBuilder();
for (int i = 0; i < asd.Count; i += 3)
sb1.AppendLine(string.Join(" ", asd.Skip(i).Take(3)));
Console.WriteLine(sb1.ToString());
sw2.Stop();
var sw3 = new Stopwatch();
sw3.Start();
var sb2 = new StringBuilder();
int counter = 0;
string output = "";
foreach (int value in asd)
{
counter++;
if (counter % 3 == 0)
{
output += value;
sb2.AppendLine(output);
output = string.Empty;
}
else
output += value + " ";
}
Console.WriteLine(sb2.ToString());
sw3.Stop();
var sw4 = new Stopwatch();
sw4.Start();
var sb3 = new StringBuilder();
for (int i = 0; i <asd.Count / 3; i++)
{
int index = i * 3;
sb3.AppendFormat("{0} {1} {2}", asd[index], asd[index + 1], asd[index + 2]);
sb3.AppendLine();
}
Console.WriteLine(sb3.ToString());
sw4.Stop();
Console.WriteLine("MySolution: " + sw1.ElapsedMilliseconds);
Console.WriteLine("Mong Zhu's Solution: " + sw2.ElapsedMilliseconds);
Console.WriteLine("Wojtek's Solution: " + sw3.ElapsedMilliseconds);
Console.WriteLine("Nirmal Subedi's Solution: " + sw4.ElapsedMilliseconds);
Console.ReadKey();
Now You have your code pasted. Anyway, I created the sample to print the 3x3 matrix.
class Program
{
static void Main(string[] args)
{
StringBuilder stringBuilder = new StringBuilder();
List<int> numbers = new List<int>() {1,2,3,4,5,6,7,8,9 };
for (int i = 0; i <3; i++)
{
int index = i * 3;
stringBuilder.AppendFormat("{0}{1}{2}", numbers[index], numbers[index + 1], numbers[index + 2]);
stringBuilder.AppendLine();
}
Console.Write(stringBuilder.ToString());
Console.ReadLine();
}
}
Is that what you meant?
string output = string.Empty;
List<int> myList = new List<int>() { 1, 2, 3, 4, 5, 6, 7, 8, 9 };
int counter = 0;
foreach (int value in myList)
{
output += value.ToString();
counter++;
if (counter % 3 == 0)
{
Console.WriteLine(output);
output = string.Empty;
}
}
This question already has answers here:
How to generate all subsets of a given size?
(5 answers)
Closed 6 years ago.
Looking for an efficient way to determine all unique subsets of a list of integers.
Say I have a List of integers containing 340 integers. I want a List of all possible subsets (of 5 elements each). All supplied integers will be unique, and the result should not duplicate any element in it's subset. An example given an input of 1,2,3,4,5,6,7,8,9 I am looking for an output of
1,2,3,4,5
1,2,3,4,6
1,2,3,4,7
1,2,3,4,8
1,2,3,4,9
1,2,3,5,6
1,2,3,5,7
...
I must do this in CSharp. can this be done in LINQ?
I have answered a several combinatorial questions, and everywhere I use a variation of a non recursive allocation free algorithm. For this case, it looks like this:
public static class Algorithms
{
public static IEnumerable<T[]> GetCombinations<T>(this T[] input, int N)
{
var result = new T[N];
var indices = new int[N];
for (int pos = 0, index = 0; ;)
{
for (; pos < N; pos++, index++)
{
indices[pos] = index;
result[pos] = input[index];
}
yield return result;
do
{
if (pos == 0) yield break;
index = indices[--pos] + 1;
}
while (index > input.Length - N + pos);
}
}
}
As in the other implementations, the method yields one and the same internal buffer, which is useful when you need just to iterate and process the resulting set once. If you need to store the combinations, then you need to clone the returned array before storing it. Here is a sample usage like in your scenario:
var input = Enumerable.Range(1, 20);
var result = input
.Distinct()
.ToArray()
.GetCombinations(5)
.Select(c => (int[])c.Clone())
.ToList();
UPDATE: The GetCombinations method basically emulates N nested loops like this (in pseudo code):
for (int i0 = 0; i0 <= input.Length - N; i0++)
for (int i1 = i0 + 1; i1 <= input.Length - N + 1; i1++)
for (int i2 = i1 + 1; i2 <= input.Length - N + 2; i2++)
...
for (int iN-1 = iN-2 + 1; iN-1 <= input.Length - 1; iN-1++)
yield { input[i0], input[i1], input[i2], ..., input[iN-1] }
In case of a tractable set of 9 elements (or max 25-30) and subset of 5, the code could be based on a recursive function
static void Main(string[] args)
{
foreach (var item in ListPerm())
{
Console.WriteLine(String.Join(",", item));
}
Console.Read();
}
private static List<List<int>> ListPerm(HashSet<int> mySet = null, int deep = 5)
{
if (mySet == null)
{
mySet = initSet(8);
}
if (deep <= 0)
{
return Enumerable.Empty<List<int>>().ToList();
}
List<List<int>> all = new List<List<int>>();
for (int i = 0; i < mySet.Count - deep + 1; i++)
{
if (deep == 1)
{
var list = new List<int>() { mySet.ElementAt(i) };
all.Add(list);
}
foreach (var item in ListPerm(new HashSet<int>(mySet.Skip(i+1)), deep - 1))
{
var list = new List<int>() { mySet.ElementAt(i) };
list.AddRange(item);
all.Add(list);
}
}
return all;
}
private static HashSet<int> initSet(int lenght)
{
HashSet<int> ret = new HashSet<int>();
for (int i = 0; i < lenght; i++)
{
ret.Add(i * 1 + 1); // just an example...
};
return ret;
}
Reengineering
Now, let me optimize the above code into a more performant function, that takes 3.2 seconds to get the combinations of 8 integers out of 30 on my standard laptop.
private static int[][] ListPerm(int[] mySet, int deep)
{
var all = new List<int[]>();
if (deep == 1)
{
return mySet.Select(x => new int[] { x }).ToArray();
}
else
{
var mySubSet = new int[mySet.Length - 1];
Array.Copy(mySet, 1, mySubSet, 0, mySubSet.Length);
var perm1st = ListPerm(mySubSet, deep - 1);
for (int i = 0; i < mySet.Length - deep + 1; i++)
{
var permn = perm1st.Select(x =>
{
var z = new int[x.Length + 1];
z[0] = mySet[i];
x.CopyTo(z, 1);
return z;
}
);
all.AddRange(permn);
int start = Array.FindIndex(perm1st, item => item[0] != mySet[i + 1]);
if (start > 0)
{
var temp_cpy = new int[perm1st.Length - start][];
Array.Copy(perm1st, start, temp_cpy, 0, temp_cpy.Length);
perm1st = temp_cpy;
}
}
}
return all.ToArray();
}
Benchmark
Here it is a comparison of Ivan's, my and the community wiki algorithms for the combinations of 5 ints in 20.
Results
wiki perm: 00:00:00.0950055
writing wiki perm: 00:00:00.0460026
Ivan perm: 00:00:00.0400023
writing Ivan perm: 00:00:00.0260015
my perm: 00:00:00.0110006
writing my perm: 00:00:00.0300017
Test Code
var input = Enumerable.Range(1, 20);
int deep = 5;
var start = DateTime.Now;
var wiki = Algorithms.Combinations(input, deep).ToArray();
Console.WriteLine("wiki perm: {0}", DateTime.Now - start);
start = DateTime.Now;
StreamWriter sw0 = new StreamWriter(#"C:\dev\SO\Algo\perm0.txt", false);
foreach (var item in wiki)
{
sw0.WriteLine(String.Join(",", item));
}
sw0.Close();
Console.WriteLine("writing wiki perm: {0}", DateTime.Now - start);
start = DateTime.Now;
start = DateTime.Now;
var result = input
.Distinct()
.ToArray()
.GetCombinations(deep)
.Select(c => (int[])c.Clone())
.ToList();
Console.WriteLine("Ivan perm: {0}", DateTime.Now - start);
start = DateTime.Now;
StreamWriter sw1 = new StreamWriter(#"C:\dev\SO\Algo\perm1.txt", false);
foreach (var item in result)
{
sw1.WriteLine(String.Join(",", item));
}
sw1.Close();
Console.WriteLine("writing Ivan perm: {0}", DateTime.Now - start);
start = DateTime.Now;
var myPerm = ListPermSO(input.ToArray(), deep);
Console.WriteLine("my perm: {0}", DateTime.Now - start);
start = DateTime.Now;
StreamWriter sw2 = new StreamWriter(#"C:\dev\SO\Algo\perm2.txt", false);
foreach (var item in myPerm)
{
sw2.WriteLine(String.Join(",", item));
}
sw2.Close();
Console.WriteLine("writing my perm: {0}", DateTime.Now - start);
Console.Read();
I wrote an algorithm to find length of longest increasing sequence in an array.
The algorithm has an array m which will contain the sequence but in some conditions, it doesn't contain the exact sequence. So in such case, I record the index and value which needs to be changed.
This algorithm is n(log n)
Now, to find the actual sequence, I loop through the array m and replace the value recorded in another array. Will my algorithm now still have the complexity if n(log n) ?
Below is the code in C#:
int[] b = { 1, 8, 5, 3, 7, 2, 9 };
int k = 1;
int i = 1;
int N = b.Length;
List<int> m = new List<int>();
int[] lup = new int[b.Length];
m.Add(0);
m.Add(b[0]);
lup[0] = 0;
while (i < N)
{
if (b[i] >= m[k])
{
k = k + 1;
m.Add(b[i]);
}
else
{
if (b[i] < m[1])
{
m[1] = b[i];
}
else
{
int j;
j = Binary_Search(m, b[i], m.Count - 1);
//if the item to be replaced was not the last element, record it
if (m[j] > b[i] && j != k)
{
lup[j] = m[j];
}
m[j] = b[i];
}
}
i = i + 1;
}
Console.WriteLine("The Input Sequence is : " + string.Join("\t", b));
Console.WriteLine("Length of Longest Up Sequence is : " + k.ToString());
List<int> result = new List<int>();
// create result based on m and lup
//DOES THIS LOOP EFFECT PERFORMANCE??
for(int x = 1; x < m.Count; x++)
{
if (lup[x] == 0)
{
result.Add(m[x]);
}
else
{
result.Add(lup[x]);
}
}
Your intuition is correct. Adding this loop is n*(log(n)+1) so it's still n*log(n).
I have this sequence 1,2,3,4,5,6,8,10,11
Expected output is 1-6,8,10-11
This problem is about formatting the sequence in easy readable form
I tried with c# and used many if & else.
Interviewer said, there is some simple algorithm to do this.
I have no idea how to achive this very simple.
Also for 1,2,3 i shown 1-3. They said its wrong!.
Is there any design pattern(interpreter) involved in this logic?
Here is one way of doing it:
int[] numbers = { 1, 2, 3, 4, 5, 6, 8, 10, 11 };
int start, end;
for (int i = 0; i < numbers.Length; i++)
{
start = numbers[i];
while (i < numbers.Length - 1 && numbers[i] + 1 == numbers[i + 1])
i++;
end = numbers[i];
if(start == end)
Console.WriteLine(start);
else
Console.WriteLine(start + " - " + end);
}
This will display subsequent numbers that grow incrementally as range. Numbers that are not increasing linearly are not written as part of a range.
Here is another version of the first approach, it utilizes the same for loop to iterate on range:
int temp = numbers[0], start, end;
for (int i = 0; i < numbers.Length; i++)
{
start = temp;
if (i < numbers.Length - 1 )
// if subsequent numbers are incremental loop further
if (numbers[i] + 1 == numbers[i + 1])
continue;
// if they are not, number at index i + 1 is a new 'start' for the next iteration
else
temp = numbers[i + 1];
end = numbers[i];
if (start == end)
Console.WriteLine(start);
else
Console.WriteLine(start + " - " + end);
}
A simple implementation in C# could look like this:
public string Format(IEnumerable<int> input)
{
var result = string.Empty;
var previous = -1;
var start = -1;
var first = true;
foreach(var i in input)
{
if(start == -1)
start = i;
else if(previous + 1 != i)
{
result += FormatRange(start, previous, first);
first = false;
start = i;
}
previous = i;
}
if(start != -1)
result += FormatRange(start, previous, first);
return result;
}
public string FormatRange(int start, int end, bool isFirst)
{
var result = string.Empty;
if(!isFirst)
result += ", ";
if(start == end)
result += start;
else
result += string.Format("{0}-{1}", start, end);
return result;
}
This will also output 1-3 for the input 1,2,3, which is perfectly valid. Without a specification what the output should be instead it's impossible to answer that part.
Probably not a suitable answer for an interview question, but using LINQ is another way to solve this.
int[] numbers = { 1, 2, 3, 4, 5, 6, 8, 10, 11 };
var remains = numbers.AsEnumerable();
while (remains.Any())
{
int first = remains.First();
int last = remains.TakeWhile((x, i) => x - first == i).Last();
remains = remains.Skip(last - first + 1);
Console.Write(first + (first == last ? "" : "-" + last) + (remains.Any() ? "," : Environment.NewLine));
}
The following groups consecutive integers, and outputs a string for each group. However, it also allows you to specify the minimum length of group which you want to hyphenate; anything less will just give you the individual numbers. Thus if you only want to hyphenate groups of 4 or more, you can pass in 4; if you want to hyphenate pairs, you can pass in 2. (I'd want to use 3 myself, but I can't tell what they want.)
It also doesn't keep any collections of numbers as it goes along, because you don't need to.
Method:
static IEnumerable<string> Group(IEnumerable<int> input, int minLength)
{
int currentStart = int.MinValue;
int currentLength = 0;
foreach (int c in input)
{
if (currentLength > 0)
if (currentStart + currentLength == c)
currentLength++;
else
{
if (currentLength >= minLength)
yield return string.Format("{0}-{1}",
currentStart, currentStart + currentLength - 1);
else
for (int i = currentStart; i < currentStart + currentLength; i++)
yield return i.ToString();
currentStart = c;
currentLength = 1;
}
else
{
currentStart = c;
currentLength = 1;
}
}
if (currentLength >= minLength)
yield return string.Format("{0}-{1}",
currentStart, currentStart + currentLength + 1);
else
for (int i = currentStart; i < currentStart + currentLength; i++)
yield return i.ToString();
}
Usage:
int minCount = 3;
int[] input = new[] { 1, 2, 3, 4, 5, 6, 8, 10, 11 };
Console.WriteLine(String.Join(",", Group(input, minCount)));
Java code:
int[] arr = {1,2,3,4,5,6,8,10,11};
int start = arr[0], last = arr[0];
String output = "";
for (int i = 1; i <= arr.length; i++)
{
if (i == arr.length || arr[i] != last+1)
{
if (output.length() != 0)
output += ",";
if (start == last)
output += start;
else
output += start + "-" + last;
if (i != arr.length)
start = last = arr[i];
}
else
last = arr[i];
}
System.out.println(output);
Heres my best attempt. Not clever, but simple enough to satisfy that requirement I believe. I'm still pretty confused as to why "1-3" was wrong though....
var numbers = new int[] { 1, 2, 3, 4, 5, 6, 8, 10, 11, 12 };
var groups = new Dictionary<int, int>();
groups.Add(numbers.First(), numbers.First());
foreach (var num in numbers.Skip(1))
{
var grp = groups.Last();
if (grp.Value + 1 == num)
{
groups[grp.Key] = num;
}
else
{
groups.Add(num, num);
}
}
var output = string.Join(",", groups.Select(grp => (grp.Key == grp.Value) ? grp.Value.ToString() : grp.Key.ToString() + "-" + grp.Value.ToString()));
Note: of course using the dictionary and linq etc is completely unnecessary (and way too specific for an answer requiring an algorithm), but I thought it highlighted the grouping aspect of the problem nicely
This is no valid C# code but to show the Idea.
Sort the list from Min to Max then do this:
For i = Min to Max
{
if i < MaxFound
continue;
int step = 1;
Output = i;
while Found(i + Step)
{
Step++;
MaxFound = i + Step;
}
if i < MaxFound
Output = (i + "-" + MaxFound);
Output += ", ";
}
Here is one of the approach:
public static void main(String[] args) {
print(1, 2, 3, 4, 5, 7, 9, 10, 12);
}
public static void print(int ... nums) {
System.out.print(nums[0]);
int idx = 1;
for(int i = 1; i < nums.length; i++, idx++) {
if(nums[i] - nums[i - 1] != 1) {
if(idx > 1) {
System.out.print(" - " + nums[i - 1]);
}
System.out.print(", " + nums[i]);
idx = 0;
}
}
if(idx > 1)
System.out.println(" - " + nums[nums.length - 1]);
}
Here's a Haskell version:
import Data.List
parseRange [] = ""
parseRange n =
let range = takeWhile (\x -> isInfixOf [x,x+1] n) n
in if not (null range)
then show (head range) ++ "-" ++ show (last range + 1)
++ (if length (tail n) > 1 then "," else "")
++ parseRange (drop (length range + 1) n)
else show (head n) ++ (if null (tail n) then "" else ",")
++ parseRange (drop 1 n)
Output:
*Main> parseRange [1,2,3,4,5,6,8,10,11]
"1-6,8,10-11"
And a way to do it with fold in F# - just for fun.
let parseRange numbers =
numbers
|> Seq.fold
(fun list n ->
match list with
|(a,b) :: tail when b+1 = n -> (a, n) :: tail
|_ -> (n,n) :: list) []
|> List.rev
|> Seq.map (fun (a,b) -> if a = b then sprintf "%i" a else sprintf "%i-%i" a b)
|> String.concat ","