Iterate all possible variants of a sequence - algorithm

I have a sequence of n letters: e.g. A B C A D , or A B F A
What I want is to get every possible variant with a comma between the letters.
i.e. for A B F A
A B F A
A,B F A
A B,F A
A B F,A
A,B,F A
A B,F,A
A,B F,A
A,B,F,A
Can anyone recommend a good algorithm for doing this? Language not important.

Simple solution to use binary array to represent if there is a comma or not.
A B F A contains three positions where comma may be (AB, BF, FA)
That means if you create 3-element array and try every possible combination of 0 and 1 you'll get the desired result. 000, 001, 010, 011, 100, 101, 110, 111
Simple program in java prints all binary permutation for n bits:
String s = "ABFA";
int bits = s.length() - 1;
int lastNumber = (int)Math.pow(2, bits);
System.out.println(lastNumber);
for (int i = 0; i < lastNumber; i++) {
System.out.println(completeZeros(Integer.toString(i, 2), bits));
}
static String completeZeros(String s, int bits) {
String result = s;
for (int i = 0; i < bits - s.length(); i++) {
result = "0" + result;
}
return result;
}
To apply binary permutation "010" to string "ABFA" use next function:
static String applyBinary(String s, String binary) {
String result = "" + s.charAt(0);
for (int i = 0; i < binary.length(); i++) {
if (binary.charAt(i) == '1') result += ", ";
result += s.charAt(i + 1);
}
return result;
}
The output is:
ABFA
ABF, A
AB, FA
AB, F, A
A, BFA
A, BF, A
A, B, FA
A, B, F, A

Use binary system for this task.
1 means comma is present, 0 means comma is not present. Each position in number informs about presence of another comma. For example for AFA:
00 : A F A
01 : A F,A
10 : A,F A
11 : A,F,A
Numbers must be taken from range [0 .. (n-1)^2-1]

Here's a simple JavaScript demonstration.
var str = "ABFA";
function algo(str) {
var result = [];
var n = str.length;
var total = Math.pow(n-1, 2) - 1;
for(var mask = 0; mask < total; mask++) {
var bits = mask;
var newstr = "";
for(var i=0; i<str.length - 1; i++, bits>>>=1) {
var hasComma = (bits & 1) == 1;
newstr += str.charAt(i);
newstr += (hasComma ? "," : " ");
}
newstr += str.charAt(str.length - 1);
result.push(newstr);
}
return result;
}
algo(str);
You calculate the total number of combinations "total"
You count up to that number "mask"
You use the binary bits of your counter "bits" to add commas

There are two approaches to this problem.
1.Recursive ( Start = "printAllComb ( s , "" , 0 );" )
printAllComb(string s, string const, int i)
{
if ( i == s.length() )
print const
printAllComb(s,const+string.at(i)+',',i+1);
printAllComb(s,const+string.at(i),i+1);
}
2.Dynamic Programming
char binaryS[s.length]="0000";
//Basically no. of zeros = no. of Alphabets in the string
//define a function AddOne() which adds 1 to the character representation
//AddOne() modifies the character array such that it stays in the bit representation
//Characters because to save space
while ( All the bits are not one )
{
for ( int i=0; i<s.length(); i++ )
{
print s.at(i)
if ( binaryS.at(i) == '1' )
print ","
}
print "\n"
AddOne();
}

Related

Algorithm - find all permutations of string a in string b

Say we have
string a = "abc"
string b = "abcdcabaabccbaa"
Find location of all permutations of a in b. I am trying to find an effective algorithm for this.
Pseudo code:
sort string a // O(a loga)
for windows of length a in b // O(b)?
sort that window of b // O(~a loga)?
compare to a
if equal
save the index
So would this be a correct algorithm? Run time would be around O(aloga + ba loga) ~= O(a loga b)? How efficient would this be? Possibly way to reduce to O(a*b) or better?
sorting is very expensive, and doesn't use the fact you move along b with a sliding window.
I would use a comparison method that is location agnostic (since any permutation is valid) - assign each letter a prime number, and each string will be the multiplication of its letter values.
this way, as you go over b, each step requires just dividing by the letter you remove from he left, and multiplying with the next letter.
You also need to convince yourself that this indeed matches uniquely for each string and covers all permutations - this comes from the uniqueness of prime decomposition. Also note that on larger strings the numbers get big so you may need some library for large numbers
There is no need to hash, you can just count frequencies on your sliding window, and check if it matches. Assuming the size of your alphabet is s, you get a very simple O(s(n + m)) algorithm.
// a = [1 .. m] and b = [1 .. n] are the input
cnta = [1 .. s] array initialized to 0
cntb = [1 .. s] array initialized to 0
// nb_matches = the number of i s.t. cnta[i] = cntb[i]
// thus the current subword = a iff. nb_matches = s
nb_matches = s
for i = 1 to m:
if cntb[a[i]] = 0: nb_matches -= 1
cntb[a[i]] += 1
ans = 0
for i = 1 to n:
if cntb[b[i]] = cnta[b[i]]: nb_matches -= 1
cntb[b[i]] += 1
if nb_matches = s: ans += 1
if cntb[b[i]] = cnta[b[i]]: nb_matches += 1
if i - m + 1 >= 1:
if cntb[b[i - m + 1]] = cnta[b[i - m + 1]]: nb_matches -= 1
cntb[b[i - m + 1]] += 1
if cntb[b[i - m + 1]] = cnta[b[i - m + 1]]: nb_matches += 1
cntb[b[i - m + 1]] -= 1
return ans
Write a function strcount() to count the number of occurrences of character ch in a string or sub-sring str.
Then just pass through the search string.
for(i=0;i<haystacklenN-NeedleN+1;i++)
{
for(j=0;j<needleN;j++)
if(strcount(haystack + i, Nneedle, needle[j]) != strcount(needles, needlesN, needle[j])
break
}
if(j == needleN)
/* found a permuatation */
Below is my solution. The space complexity is just O(a + b), and the running time (if I can calculate correctly..) is O(b*a), as for each character in b, we may do a recursion a levels deep.
md5's answer is a good one and will be faster!!
public class FindPermutations {
public static void main(String[] args) {
System.out.println(numPerms(new String("xacxzaa"),
new String("fxaazxacaaxzoecazxaxaz")));
System.out.println(numPerms(new String("ABCD"),
new String("BACDGABCDA")));
System.out.println(numPerms(new String("AABA"),
new String("AAABABAA")));
// prints 4, then 3, then 3
}
public static int numPerms(final String a, final String b) {
int sum = 0;
for (int i = 0; i < b.length(); i++) {
if (permPresent(a, b.substring(i))) {
sum++;
}
}
return sum;
}
// is a permutation of a present at the start of b?
public static boolean permPresent(final String a, final String b) {
if (a.isEmpty()) {
return true;
}
if (b.isEmpty()) {
return false;
}
final char first = b.charAt(0);
if (a.contains(b.substring(0, 1))) {
// super ugly, but removes first from a
return permPresent(a.substring(0, a.indexOf(first)) + a.substring(a.indexOf(first)+1, a.length()),
b.substring(1));
}
return false;
}
}
For searchability's sake, I arrive on this page afer looking for other solutions to compare mine to, with the problem originating from watching this clip: https://www.hackerrank.com/domains/tutorials/cracking-the-coding-interview. The original problem statement was something like 'find all permutations of s in b'.
Use 2 hash tables and with a sliding window of size = length of smaller string:
int premutations_of_B_in_A(string large, string small) {
unordered_map<char, int> characters_in_large;
unordered_map<char, int> characters_in_small;
int ans = 0;
for (char c : small) {
characters_in_small[c]++;
}
for (int i = 0; i < small.length(); i++) {
characters_in_large[large[i]]++;
ans += (characters_in_small == characters_in_large);
}
for (int i = small.length(); i < large.length(); i++) {
characters_in_large[large[i]]++;
if (characters_in_large[large[i - small.length()]]-- == 1)
characters_in_large.erase(large[i - small.length()]);
ans += (characters_in_small == characters_in_large);
}
return ans;
}
This is almost solution but will help you to count occurrences of permutations of small strings into larger string
made for only lower case chars
This solution having --
Time Complexity - O(L)
where L is length of large input provided to problem, the exact would be to include 26 too for every char present in Large array but by ignoring constant terms, I will solely stand for this.
Space Complexity - O(1)
because 26 is also constant and independent of how large input would be.
int findAllPermutations(string small, string larger) {
int freqSmall[26] = {0};
//window size
int n = small.length();
//to return
int finalAns = 0;
for (char a : small) {
freqSmall[a - 97]++;
}
int freqlarger[26]={0};
int count = 0;
int j = 0;
for (int i = 0; larger[i] != '\0'; i++) {
freqlarger[larger[i] - 97]++;
count++;
if (count == n) {
count = 0;
int i;
for (i = 0; i < 26; i++) {
if (freqlarger[i] != freqSmall[i]) {
break;
}
}
if (i == 26) {
finalAns++;
}
freqlarger[larger[j] - 97]--;
j++;
}
}
return finalAns;
}
int main() {
string s, t;
cin >> s >> t;
cout << findAllPermutations(s, t) << endl;
return 0;
}

Clean algorithm to get Excel column letters from column index [duplicate]

How do you convert a numerical number to an Excel column name in C# without using automation getting the value directly from Excel.
Excel 2007 has a possible range of 1 to 16384, which is the number of columns that it supports. The resulting values should be in the form of excel column names, e.g. A, AA, AAA etc.
Here's how I do it:
private string GetExcelColumnName(int columnNumber)
{
string columnName = "";
while (columnNumber > 0)
{
int modulo = (columnNumber - 1) % 26;
columnName = Convert.ToChar('A' + modulo) + columnName;
columnNumber = (columnNumber - modulo) / 26;
}
return columnName;
}
If anyone needs to do this in Excel without VBA, here is a way:
=SUBSTITUTE(ADDRESS(1;colNum;4);"1";"")
where colNum is the column number
And in VBA:
Function GetColumnName(colNum As Integer) As String
Dim d As Integer
Dim m As Integer
Dim name As String
d = colNum
name = ""
Do While (d > 0)
m = (d - 1) Mod 26
name = Chr(65 + m) + name
d = Int((d - m) / 26)
Loop
GetColumnName = name
End Function
You might need conversion both ways, e.g from Excel column adress like AAZ to integer and from any integer to Excel. The two methods below will do just that. Assumes 1 based indexing, first element in your "arrays" are element number 1.
No limits on size here, so you can use adresses like ERROR and that would be column number 2613824 ...
public static string ColumnAdress(int col)
{
if (col <= 26) {
return Convert.ToChar(col + 64).ToString();
}
int div = col / 26;
int mod = col % 26;
if (mod == 0) {mod = 26;div--;}
return ColumnAdress(div) + ColumnAdress(mod);
}
public static int ColumnNumber(string colAdress)
{
int[] digits = new int[colAdress.Length];
for (int i = 0; i < colAdress.Length; ++i)
{
digits[i] = Convert.ToInt32(colAdress[i]) - 64;
}
int mul=1;int res=0;
for (int pos = digits.Length - 1; pos >= 0; --pos)
{
res += digits[pos] * mul;
mul *= 26;
}
return res;
}
Sorry, this is Python instead of C#, but at least the results are correct:
def ColIdxToXlName(idx):
if idx < 1:
raise ValueError("Index is too small")
result = ""
while True:
if idx > 26:
idx, r = divmod(idx - 1, 26)
result = chr(r + ord('A')) + result
else:
return chr(idx + ord('A') - 1) + result
for i in xrange(1, 1024):
print "%4d : %s" % (i, ColIdxToXlName(i))
I discovered an error in my first post, so I decided to sit down and do the the math. What I found is that the number system used to identify Excel columns is not a base 26 system, as another person posted. Consider the following in base 10. You can also do this with the letters of the alphabet.
Space:.........................S1, S2, S3 : S1, S2, S3
....................................0, 00, 000 :.. A, AA, AAA
....................................1, 01, 001 :.. B, AB, AAB
.................................... …, …, … :.. …, …, …
....................................9, 99, 999 :.. Z, ZZ, ZZZ
Total states in space: 10, 100, 1000 : 26, 676, 17576
Total States:...............1110................18278
Excel numbers columns in the individual alphabetical spaces using base 26. You can see that in general, the state space progression is a, a^2, a^3, … for some base a, and the total number of states is a + a^2 + a^3 + … .
Suppose you want to find the total number of states A in the first N spaces. The formula for doing so is A = (a)(a^N - 1 )/(a-1). This is important because we need to find the space N that corresponds to our index K. If I want to find out where K lies in the number system I need to replace A with K and solve for N. The solution is N = log{base a} (A (a-1)/a +1). If I use the example of a = 10 and K = 192, I know that N = 2.23804… . This tells me that K lies at the beginning of the third space since it is a little greater than two.
The next step is to find exactly how far in the current space we are. To find this, subtract from K the A generated using the floor of N. In this example, the floor of N is two. So, A = (10)(10^2 – 1)/(10-1) = 110, as is expected when you combine the states of the first two spaces. This needs to be subtracted from K because these first 110 states would have already been accounted for in the first two spaces. This leaves us with 82 states. So, in this number system, the representation of 192 in base 10 is 082.
The C# code using a base index of zero is
private string ExcelColumnIndexToName(int Index)
{
string range = string.Empty;
if (Index < 0 ) return range;
int a = 26;
int x = (int)Math.Floor(Math.Log((Index) * (a - 1) / a + 1, a));
Index -= (int)(Math.Pow(a, x) - 1) * a / (a - 1);
for (int i = x+1; Index + i > 0; i--)
{
range = ((char)(65 + Index % a)).ToString() + range;
Index /= a;
}
return range;
}
//Old Post
A zero-based solution in C#.
private string ExcelColumnIndexToName(int Index)
{
string range = "";
if (Index < 0 ) return range;
for(int i=1;Index + i > 0;i=0)
{
range = ((char)(65 + Index % 26)).ToString() + range;
Index /= 26;
}
if (range.Length > 1) range = ((char)((int)range[0] - 1)).ToString() + range.Substring(1);
return range;
}
This answer is in javaScript:
function getCharFromNumber(columnNumber){
var dividend = columnNumber;
var columnName = "";
var modulo;
while (dividend > 0)
{
modulo = (dividend - 1) % 26;
columnName = String.fromCharCode(65 + modulo).toString() + columnName;
dividend = parseInt((dividend - modulo) / 26);
}
return columnName;
}
Easy with recursion.
public static string GetStandardExcelColumnName(int columnNumberOneBased)
{
int baseValue = Convert.ToInt32('A');
int columnNumberZeroBased = columnNumberOneBased - 1;
string ret = "";
if (columnNumberOneBased > 26)
{
ret = GetStandardExcelColumnName(columnNumberZeroBased / 26) ;
}
return ret + Convert.ToChar(baseValue + (columnNumberZeroBased % 26) );
}
I'm surprised all of the solutions so far contain either iteration or recursion.
Here's my solution that runs in constant time (no loops). This solution works for all possible Excel columns and checks that the input can be turned into an Excel column. Possible columns are in the range [A, XFD] or [1, 16384]. (This is dependent on your version of Excel)
private static string Turn(uint col)
{
if (col < 1 || col > 16384) //Excel columns are one-based (one = 'A')
throw new ArgumentException("col must be >= 1 and <= 16384");
if (col <= 26) //one character
return ((char)(col + 'A' - 1)).ToString();
else if (col <= 702) //two characters
{
char firstChar = (char)((int)((col - 1) / 26) + 'A' - 1);
char secondChar = (char)(col % 26 + 'A' - 1);
if (secondChar == '#') //Excel is one-based, but modulo operations are zero-based
secondChar = 'Z'; //convert one-based to zero-based
return string.Format("{0}{1}", firstChar, secondChar);
}
else //three characters
{
char firstChar = (char)((int)((col - 1) / 702) + 'A' - 1);
char secondChar = (char)((col - 1) / 26 % 26 + 'A' - 1);
char thirdChar = (char)(col % 26 + 'A' - 1);
if (thirdChar == '#') //Excel is one-based, but modulo operations are zero-based
thirdChar = 'Z'; //convert one-based to zero-based
return string.Format("{0}{1}{2}", firstChar, secondChar, thirdChar);
}
}
Same implementation in Java
public String getExcelColumnName (int columnNumber)
{
int dividend = columnNumber;
int i;
String columnName = "";
int modulo;
while (dividend > 0)
{
modulo = (dividend - 1) % 26;
i = 65 + modulo;
columnName = new Character((char)i).toString() + columnName;
dividend = (int)((dividend - modulo) / 26);
}
return columnName;
}
int nCol = 127;
string sChars = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
string sCol = "";
while (nCol >= 26)
{
int nChar = nCol % 26;
nCol = (nCol - nChar) / 26;
// You could do some trick with using nChar as offset from 'A', but I am lazy to do it right now.
sCol = sChars[nChar] + sCol;
}
sCol = sChars[nCol] + sCol;
Update: Peter's comment is right. That's what I get for writing code in the browser. :-) My solution was not compiling, it was missing the left-most letter and it was building the string in reverse order - all now fixed.
Bugs aside, the algorithm is basically converting a number from base 10 to base 26.
Update 2: Joel Coehoorn is right - the code above will return AB for 27. If it was real base 26 number, AA would be equal to A and the next number after Z would be BA.
int nCol = 127;
string sChars = "0ABCDEFGHIJKLMNOPQRSTUVWXYZ";
string sCol = "";
while (nCol > 26)
{
int nChar = nCol % 26;
if (nChar == 0)
nChar = 26;
nCol = (nCol - nChar) / 26;
sCol = sChars[nChar] + sCol;
}
if (nCol != 0)
sCol = sChars[nCol] + sCol;
..And converted to php:
function GetExcelColumnName($columnNumber) {
$columnName = '';
while ($columnNumber > 0) {
$modulo = ($columnNumber - 1) % 26;
$columnName = chr(65 + $modulo) . $columnName;
$columnNumber = (int)(($columnNumber - $modulo) / 26);
}
return $columnName;
}
Just throwing in a simple two-line C# implementation using recursion, because all the answers here seem far more complicated than necessary.
/// <summary>
/// Gets the column letter(s) corresponding to the given column number.
/// </summary>
/// <param name="column">The one-based column index. Must be greater than zero.</param>
/// <returns>The desired column letter, or an empty string if the column number was invalid.</returns>
public static string GetColumnLetter(int column) {
if (column < 1) return String.Empty;
return GetColumnLetter((column - 1) / 26) + (char)('A' + (column - 1) % 26);
}
Although there are already a bunch of valid answers1, none get into the theory behind it.
Excel column names are bijective base-26 representations of their number. This is quite different than an ordinary base 26 (there is no leading zero), and I really recommend reading the Wikipedia entry to grasp the differences. For example, the decimal value 702 (decomposed in 26*26 + 26) is represented in "ordinary" base 26 by 110 (i.e. 1x26^2 + 1x26^1 + 0x26^0) and in bijective base-26 by ZZ (i.e. 26x26^1 + 26x26^0).
Differences aside, bijective numeration is a positional notation, and as such we can perform conversions using an iterative (or recursive) algorithm which on each iteration finds the digit of the next position (similarly to an ordinary base conversion algorithm).
The general formula to get the digit at the last position (the one indexed 0) of the bijective base-k representation of a decimal number m is (f being the ceiling function minus 1):
m - (f(m / k) * k)
The digit at the next position (i.e. the one indexed 1) is found by applying the same formula to the result of f(m / k). We know that for the last digit (i.e. the one with the highest index) f(m / k) is 0.
This forms the basis for an iteration that finds each successive digit in bijective base-k of a decimal number. In pseudo-code it would look like this (digit() maps a decimal integer to its representation in the bijective base -- e.g. digit(1) would return A in bijective base-26):
fun conv(m)
q = f(m / k)
a = m - (q * k)
if (q == 0)
return digit(a)
else
return conv(q) + digit(a);
So we can translate this to C#2 to get a generic3 "conversion to bijective base-k" ToBijective() routine:
class BijectiveNumeration {
private int baseK;
private Func<int, char> getDigit;
public BijectiveNumeration(int baseK, Func<int, char> getDigit) {
this.baseK = baseK;
this.getDigit = getDigit;
}
public string ToBijective(double decimalValue) {
double q = f(decimalValue / baseK);
double a = decimalValue - (q * baseK);
return ((q > 0) ? ToBijective(q) : "") + getDigit((int)a);
}
private static double f(double i) {
return (Math.Ceiling(i) - 1);
}
}
Now for conversion to bijective base-26 (our "Excel column name" use case):
static void Main(string[] args)
{
BijectiveNumeration bijBase26 = new BijectiveNumeration(
26,
(value) => Convert.ToChar('A' + (value - 1))
);
Console.WriteLine(bijBase26.ToBijective(1)); // prints "A"
Console.WriteLine(bijBase26.ToBijective(26)); // prints "Z"
Console.WriteLine(bijBase26.ToBijective(27)); // prints "AA"
Console.WriteLine(bijBase26.ToBijective(702)); // prints "ZZ"
Console.WriteLine(bijBase26.ToBijective(16384)); // prints "XFD"
}
Excel's maximum column index is 16384 / XFD, but this code will convert any positive number.
As an added bonus, we can now easily convert to any bijective base. For example for bijective base-10:
static void Main(string[] args)
{
BijectiveNumeration bijBase10 = new BijectiveNumeration(
10,
(value) => value < 10 ? Convert.ToChar('0'+value) : 'A'
);
Console.WriteLine(bijBase10.ToBijective(1)); // prints "1"
Console.WriteLine(bijBase10.ToBijective(10)); // prints "A"
Console.WriteLine(bijBase10.ToBijective(123)); // prints "123"
Console.WriteLine(bijBase10.ToBijective(20)); // prints "1A"
Console.WriteLine(bijBase10.ToBijective(100)); // prints "9A"
Console.WriteLine(bijBase10.ToBijective(101)); // prints "A1"
Console.WriteLine(bijBase10.ToBijective(2010)); // prints "19AA"
}
1 This generic answer can eventually be reduced to the other, correct, specific answers, but I find it hard to fully grasp the logic of the solutions without the formal theory behind bijective numeration in general. It also proves its correctness nicely. Additionally, several similar questions link back to this one, some being language-agnostic or more generic. That's why I thought the addition of this answer was warranted, and that this question was a good place to put it.
2 C# disclaimer: I implemented an example in C# because this is what is asked here, but I have never learned nor used the language. I have verified it does compile and run, but please adapt it to fit the language best practices / general conventions, if necessary.
3 This example only aims to be correct and understandable ; it could and should be optimized would performance matter (e.g. with tail-recursion -- but that seems to require trampolining in C#), and made safer (e.g. by validating parameters).
I wanted to throw in my static class I use, for interoping between col index and col Label. I use a modified accepted answer for my ColumnLabel Method
public static class Extensions
{
public static string ColumnLabel(this int col)
{
var dividend = col;
var columnLabel = string.Empty;
int modulo;
while (dividend > 0)
{
modulo = (dividend - 1) % 26;
columnLabel = Convert.ToChar(65 + modulo).ToString() + columnLabel;
dividend = (int)((dividend - modulo) / 26);
}
return columnLabel;
}
public static int ColumnIndex(this string colLabel)
{
// "AD" (1 * 26^1) + (4 * 26^0) ...
var colIndex = 0;
for(int ind = 0, pow = colLabel.Count()-1; ind < colLabel.Count(); ++ind, --pow)
{
var cVal = Convert.ToInt32(colLabel[ind]) - 64; //col A is index 1
colIndex += cVal * ((int)Math.Pow(26, pow));
}
return colIndex;
}
}
Use this like...
30.ColumnLabel(); // "AD"
"AD".ColumnIndex(); // 30
private String getColumn(int c) {
String s = "";
do {
s = (char)('A' + (c % 26)) + s;
c /= 26;
} while (c-- > 0);
return s;
}
Its not exactly base 26, there is no 0 in the system. If there was, 'Z' would be followed by 'BA' not by 'AA'.
if you just want it for a cell formula without code, here's a formula for it:
IF(COLUMN()>=26,CHAR(ROUND(COLUMN()/26,1)+64)&CHAR(MOD(COLUMN(),26)+64),CHAR(COLUMN()+64))
In Delphi (Pascal):
function GetExcelColumnName(columnNumber: integer): string;
var
dividend, modulo: integer;
begin
Result := '';
dividend := columnNumber;
while dividend > 0 do begin
modulo := (dividend - 1) mod 26;
Result := Chr(65 + modulo) + Result;
dividend := (dividend - modulo) div 26;
end;
end;
A little late to the game, but here's the code I use (in C#):
private static readonly string _Alphabet = "ABCDEFGHIJKLMNOPQRSTUVWXYZ";
public static int ColumnNameParse(string value)
{
// assumes value.Length is [1,3]
// assumes value is uppercase
var digits = value.PadLeft(3).Select(x => _Alphabet.IndexOf(x));
return digits.Aggregate(0, (current, index) => (current * 26) + (index + 1));
}
In perl, for an input of 1 (A), 27 (AA), etc.
sub excel_colname {
my ($idx) = #_; # one-based column number
--$idx; # zero-based column index
my $name = "";
while ($idx >= 0) {
$name .= chr(ord("A") + ($idx % 26));
$idx = int($idx / 26) - 1;
}
return scalar reverse $name;
}
Though I am late to the game, Graham's answer is far from being optimal. Particularly, you don't have to use the modulo, call ToString() and apply (int) cast. Considering that in most cases in C# world you would start numbering from 0, here is my revision:
public static string GetColumnName(int index) // zero-based
{
const byte BASE = 'Z' - 'A' + 1;
string name = String.Empty;
do
{
name = Convert.ToChar('A' + index % BASE) + name;
index = index / BASE - 1;
}
while (index >= 0);
return name;
}
More than 30 solutions already, but here's my one-line C# solution...
public string IntToExcelColumn(int i)
{
return ((i<16926? "" : ((char)((((i/26)-1)%26)+65)).ToString()) + (i<2730? "" : ((char)((((i/26)-1)%26)+65)).ToString()) + (i<26? "" : ((char)((((i/26)-1)%26)+65)).ToString()) + ((char)((i%26)+65)));
}
After looking at all the supplied Versions here, I decided to do one myself, using recursion.
Here is my vb.net Version:
Function CL(ByVal x As Integer) As String
If x >= 1 And x <= 26 Then
CL = Chr(x + 64)
Else
CL = CL((x - x Mod 26) / 26) & Chr((x Mod 26) + 1 + 64)
End If
End Function
Refining the original solution (in C#):
public static class ExcelHelper
{
private static Dictionary<UInt16, String> l_DictionaryOfColumns;
public static ExcelHelper() {
l_DictionaryOfColumns = new Dictionary<ushort, string>(256);
}
public static String GetExcelColumnName(UInt16 l_Column)
{
UInt16 l_ColumnCopy = l_Column;
String l_Chars = "0ABCDEFGHIJKLMNOPQRSTUVWXYZ";
String l_rVal = "";
UInt16 l_Char;
if (l_DictionaryOfColumns.ContainsKey(l_Column) == true)
{
l_rVal = l_DictionaryOfColumns[l_Column];
}
else
{
while (l_ColumnCopy > 26)
{
l_Char = l_ColumnCopy % 26;
if (l_Char == 0)
l_Char = 26;
l_ColumnCopy = (l_ColumnCopy - l_Char) / 26;
l_rVal = l_Chars[l_Char] + l_rVal;
}
if (l_ColumnCopy != 0)
l_rVal = l_Chars[l_ColumnCopy] + l_rVal;
l_DictionaryOfColumns.ContainsKey(l_Column) = l_rVal;
}
return l_rVal;
}
}
Here is an Actionscript version:
private var columnNumbers:Array = ['A', 'B', 'C', 'D', 'E', 'F' , 'G', 'H', 'I', 'J', 'K' ,'L','M','N','O','P','Q','R','S','T','U','V','W','X','Y','Z'];
private function getExcelColumnName(columnNumber:int) : String{
var dividend:int = columnNumber;
var columnName:String = "";
var modulo:int;
while (dividend > 0)
{
modulo = (dividend - 1) % 26;
columnName = columnNumbers[modulo] + columnName;
dividend = int((dividend - modulo) / 26);
}
return columnName;
}
JavaScript Solution
/**
* Calculate the column letter abbreviation from a 1 based index
* #param {Number} value
* #returns {string}
*/
getColumnFromIndex = function (value) {
var base = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ'.split('');
var remainder, result = "";
do {
remainder = value % 26;
result = base[(remainder || 26) - 1] + result;
value = Math.floor(value / 26);
} while (value > 0);
return result;
};
These my codes to convert specific number (index start from 1) to Excel Column.
public static string NumberToExcelColumn(uint number)
{
uint originalNumber = number;
uint numChars = 1;
while (Math.Pow(26, numChars) < number)
{
numChars++;
if (Math.Pow(26, numChars) + 26 >= number)
{
break;
}
}
string toRet = "";
uint lastValue = 0;
do
{
number -= lastValue;
double powerVal = Math.Pow(26, numChars - 1);
byte thisCharIdx = (byte)Math.Truncate((columnNumber - 1) / powerVal);
lastValue = (int)powerVal * thisCharIdx;
if (numChars - 2 >= 0)
{
double powerVal_next = Math.Pow(26, numChars - 2);
byte thisCharIdx_next = (byte)Math.Truncate((columnNumber - lastValue - 1) / powerVal_next);
int lastValue_next = (int)Math.Pow(26, numChars - 2) * thisCharIdx_next;
if (thisCharIdx_next == 0 && lastValue_next == 0 && powerVal_next == 26)
{
thisCharIdx--;
lastValue = (int)powerVal * thisCharIdx;
}
}
toRet += (char)((byte)'A' + thisCharIdx + ((numChars > 1) ? -1 : 0));
numChars--;
} while (numChars > 0);
return toRet;
}
My Unit Test:
[TestMethod]
public void Test()
{
Assert.AreEqual("A", NumberToExcelColumn(1));
Assert.AreEqual("Z", NumberToExcelColumn(26));
Assert.AreEqual("AA", NumberToExcelColumn(27));
Assert.AreEqual("AO", NumberToExcelColumn(41));
Assert.AreEqual("AZ", NumberToExcelColumn(52));
Assert.AreEqual("BA", NumberToExcelColumn(53));
Assert.AreEqual("ZZ", NumberToExcelColumn(702));
Assert.AreEqual("AAA", NumberToExcelColumn(703));
Assert.AreEqual("ABC", NumberToExcelColumn(731));
Assert.AreEqual("ACQ", NumberToExcelColumn(771));
Assert.AreEqual("AYZ", NumberToExcelColumn(1352));
Assert.AreEqual("AZA", NumberToExcelColumn(1353));
Assert.AreEqual("AZB", NumberToExcelColumn(1354));
Assert.AreEqual("BAA", NumberToExcelColumn(1379));
Assert.AreEqual("CNU", NumberToExcelColumn(2413));
Assert.AreEqual("GCM", NumberToExcelColumn(4823));
Assert.AreEqual("MSR", NumberToExcelColumn(9300));
Assert.AreEqual("OMB", NumberToExcelColumn(10480));
Assert.AreEqual("ULV", NumberToExcelColumn(14530));
Assert.AreEqual("XFD", NumberToExcelColumn(16384));
}
Sorry, this is Python instead of C#, but at least the results are correct:
def excel_column_number_to_name(column_number):
output = ""
index = column_number-1
while index >= 0:
character = chr((index%26)+ord('A'))
output = output + character
index = index/26 - 1
return output[::-1]
for i in xrange(1, 1024):
print "%4d : %s" % (i, excel_column_number_to_name(i))
Passed these test cases:
Column Number: 494286 => ABCDZ
Column Number: 27 => AA
Column Number: 52 => AZ
For what it is worth, here is Graham's code in Powershell:
function ConvertTo-ExcelColumnID {
param (
[parameter(Position = 0,
HelpMessage = "A 1-based index to convert to an excel column ID. e.g. 2 => 'B', 29 => 'AC'",
Mandatory = $true)]
[int]$index
);
[string]$result = '';
if ($index -le 0 ) {
return $result;
}
while ($index -gt 0) {
[int]$modulo = ($index - 1) % 26;
$character = [char]($modulo + [int][char]'A');
$result = $character + $result;
[int]$index = ($index - $modulo) / 26;
}
return $result;
}
Another VBA way
Public Function GetColumnName(TargetCell As Range) As String
GetColumnName = Split(CStr(TargetCell.Cells(1, 1).Address), "$")(1)
End Function
Here's my super late implementation in PHP. This one's recursive. I wrote it just before I found this post. I wanted to see if others had solved this problem already...
public function GetColumn($intNumber, $strCol = null) {
if ($intNumber > 0) {
$intRem = ($intNumber - 1) % 26;
$strCol = $this->GetColumn(intval(($intNumber - $intRem) / 26), sprintf('%s%s', chr(65 + $intRem), $strCol));
}
return $strCol;
}

minimum reduced string made up of a,b,c [duplicate]

I have a question which asks us to reduce the string as follows.
The input is a string having only A, B or C. Output must be length of
the reduced string
The string can be reduced by the following rules
If any 2 different letters are adjacent, these two letters can be
replaced by the third letter.
Eg ABA -> CA -> B . So final answer is 1 (length of reduced string)
Eg ABCCCCCCC
This doesn't become CCCCCCCC, as it can be reduced alternatively by
ABCCCCCCC->AACCCCCC->ABCCCCC->AACCCC->ABCCC->AACC->ABC->AA
as here length is 2 < (length of CCCCCCCC)
How do you go about this problem?
Thanks a lot!
To make things clear: the question states it wants the minimum length of the reduced string. So in the second example above there are 2 solutions possible, one CCCCCCCC and the other AA. So 2 is the answer as length of AA is 2 which is smaller than the length of CCCCCCCC = 8.
The way this question is phrased, there are only three distinct possibilities:
If the string has only one unique character, the length is the same as the length of the string.
2/3. If the string contains more than one unique character, the length is either 1 or 2, always (based on the layout of the characters).
Edit:
As a way of proof of concept here is some grammar and its extensions:
I should note that although this seems to me a reasonable proof for the fact that the length will reduce to either 1 or 2, I am reasonably sure that determining which of these lengths will result is not as trivial as I originally thought ( you would still have to recurse through all options to find it out)
S : A|B|C|()
S : S^
where () denotes the empty string, and s^ means any combination of the previous [A,B,C,()] characters.
Extended Grammar:
S_1 : AS^|others
S_2 : AAS^|ABS^|ACS^|others
S_3 : AAAS^|
AABS^ => ACS^ => BS^|
AACS^ => ABS^ => CS^|
ABAS^ => ACS^ => BS^|
ABBS^ => CBS^ => AS^|
ABCS^ => CCS^ | AAS^|
ACAS^ => ABS^ => CS^|
ACBS^ => AAS^ | BBS^|
ACCS^ => BCS^ => AS^|
The same thing will happen with extended grammars starting with B, and C (others). The interesting cases are where we have ACB and ABC (three distinct characters in sequence), these cases result in grammars that appear to lead to longer lengths however:
CCS^: CCAS^|CCBS^|CCCS^|
CBS^ => AS^|
CAS^ => BS^|
CCCS^|
AAS^: AAAS^|AABS^|AACS^|
ACS^ => BS^|
ABS^ => CS^|
AAAS^|
BBS^: BBAS^|BBBS^|BBCS^|
BCS^ => AS^|
BAS^ => CS^|
BBBS^|
Recursively they only lead to longer lengths when the remaining string contains their value only. However we have to remember that this case also can be simplified, since if we got to this area with say CCCS^, then we at one point previous had ABC ( or consequently CBA ). If we look back we could have made better decisions:
ABCCS^ => AACS^ => ABS^ => CS^
CBACS^ => CBBS^ => ABS^ => CS^
So in the best case at the end of the string when we make all the correct decisions we end with a remaining string of 1 character followed by 1 more character(since we are at the end). At this time if the character is the same, then we have a length of 2, if it is different, then we can reduce one last time and we end up with a length of 1.
You can generalize the result based on individual character count of string. The algo is as follows,
traverse through the string and get individual char count.
Lets say if
a = no# of a's in given string
b = no# of b's in given string
c = no# of c's in given string
then you can say that, the result will be,
if((a == 0 && b == 0 && c == 0) ||
(a == 0 && b == 0 && c != 0) ||
(a == 0 && b != 0 && c == 0) ||
(a != 0 && b == 0 && c == 0))
{
result = a+b+c;
}
else if(a != 0 && b != 0 && c != 0)
{
if((a%2 == 0 && b%2 == 0 && c%2 == 0) ||
(a%2 == 1 && b%2 == 1 && c%2 == 1))
result = 2;
else
result = 1;
}
else if((a == 0 && b != 0 && c != 0) ||
(a != 0 && b == 0 && c != 0) ||
(a != 0 && b != 0 && c == 0))
{
if(a%2 == 0 && b%2 == 0 && c%2 == 0)
result = 2;
else
result = 1;
}
I'm assuming that you are looking for the length of the shortest possible string that can be obtained after reduction.
A simple solution would be to explore all possibilities in a greedy manner and hope that it does not explode exponentially. I'm gonna write Python pseudocode here because that's easier to comprehend (at least for me ;)):
from collections import deque
def try_reduce(string):
queue = deque([string])
min_length = len(string)
while queue:
string = queue.popleft()
if len(string) < min_length:
min_length = len(string)
for i in xrange(len(string)-1):
substring = string[i:(i+2)]
if substring == "AB" or substring == "BA":
queue.append(string[:i] + "C" + string[(i+2):])
elif substring == "BC" or substring == "CB":
queue.append(string[:i] + "A" + string[(i+2):])
elif substring == "AC" or substring == "CA":
queue.append(string[:i] + "B" + string[(i+2):])
return min_length
I think the basic idea is clear: you take a queue (std::deque should be just fine), add your string into it, and then implement a simple breadth first search in the space of all possible reductions. During the search, you take the first element from the queue, take all possible substrings of it, execute all possible reductions, and push the reduced strings back to the queue. The entire space is explored when the queue becomes empty.
Let's define an automaton with the following rules (K>=0):
Incoming: A B C
Current: --------------------------
<empty> A B C
A(2K+1) A(2K+2) AB AC
A(2K+2) A(2K+3) AAB AAC
AB CA CB ABC
AAB BA ACB BC
ABC CCA AAB AAC
and all rules obtained by permutations of ABC to get the complete definition.
All input strings using a single letter are irreducible. If the input string contains at least two different letters, the final states like AB or AAB can be reduced to a single letter, and the final states like ABC can be reduced to two letters.
In the ABC case, we still have to prove that the input string can't be reduced to a single letter by another reduction sequence.
Compare two characters at a time and replace if both adjacent characters are not same. To get optimal solution, run once from start of the string and once from end of the string. Return the minimum value.
int same(char* s){
int i=0;
for(i=0;i<strlen(s)-1;i++){
if(*(s+i) == *(s+i+1))
continue;
else
return 0;
}
return 1;
}
int reduceb(char* s){
int ret = 0,a_sum=0,i=0;
int len = strlen(s);
while(1){
i=len-1;
while(i>0){
if ((*(s+i)) == (*(s+i-1))){
i--;
continue;
} else {
a_sum = (*(s+i)) + (*(s+i-1));
*(s+i-1) = SUM - a_sum;
*(s+i) = '\0';
len--;
}
i--;
}
if(same(s) == 1){
return strlen(s);
}
}
}
int reducef(char* s){
int ret = 0,a_sum=0,i=0;
int len = strlen(s);
while(1){
i=0;
while(i<len-1){
if ((*(s+i)) == (*(s+i+1))){
i++;
continue;
} else {
a_sum = (*(s+i)) + (*(s+i+1));
*(s+i) = SUM - a_sum;
int j=i+1;
for(j=i+1;j<len;j++)
*(s+j) = *(s+j+1);
len--;
}
i++;
}
if(same(s) == 1){
return strlen(s);
}
}
}
int main(){
int n,i=0,f=0,b=0;
scanf("%d",&n);
int a[n];
while(i<n){
char* str = (char*)malloc(101);
scanf("%s",str);
char* strd = strdup(str);
f = reducef(str);
b = reduceb(strd);
if( f > b)
a[i] = b;
else
a[i] = f;
free(str);
free(strd);
i++;
}
for(i=0;i<n;i++)
printf("%d\n",a[i]);
}
import java.io.*;
import java.util.*;
class StringSim{
public static void main(String args[]){
Scanner sc = new Scanner(System.in);
StringTokenizer st = new StringTokenizer(sc.nextLine(), " ");
int N = Integer.parseInt(st.nextToken());
String op = "";
for(int i=0;i<N;i++){
String str = sc.nextLine();
op = op + Count(str) + "\n";
}
System.out.println(op);
}
public static int Count( String str){
int min = Integer.MAX_VALUE;
char pre = str.charAt(0);
boolean allSame = true;
//System.out.println("str :" + str);
if(str.length() == 1){
return 1;
}
int count = 1;
for(int i=1;i<str.length();i++){
//System.out.println("pre: -"+ pre +"- char at "+i+" is : -"+ str.charAt(i)+"-");
if(pre != str.charAt(i)){
allSame = false;
char rep = (char)(('a'+'b'+'c')-(pre+str.charAt(i)));
//System.out.println("rep :" + rep);
if(str.length() == 2)
count = 1;
else if(i==1)
count = Count(rep+str.substring(2,str.length()));
else if(i == str.length()-1)
count = Count(str.substring(0,str.length()-2)+rep);
else
count = Count(str.substring(0,i-1)+rep+str.substring(i+1,str.length()));
if(min>count) min=count;
}else if(allSame){
count++;
//System.out.println("count: " + count);
}
pre = str.charAt(i);
}
//System.out.println("min: " + min);
if(allSame) return count;
return min;
}
}
Wouldn't a good start be to count which letter you have the most of and look for ways to remove it? Keep doing this until we only have one letter. We might have it many times but as long as it is the same we do not care, we are finished.
To avoid getting something like ABCCCCCCC becoming CCCCCCCC.
We remove the most popular letter:
-ABCCCCCCC
-AACCCCCC
-ABCCCCC
-AACCCC
-ABCCC
-AACC
-ABC
-AA
I disagree with the previous poster who states we must have a length of 1 or 2 - what happens if I enter the start string AAA?
import java.util.LinkedList;
import java.util.List;
import java.util.Scanner;
public class Sample {
private static char[] res = {'a', 'b', 'c'};
private char replacementChar(char a, char b) {
for(char c : res) {
if(c != a && c != b) {
return c;
}
}
throw new IllegalStateException("cannot happen. you must've mucked up the resource");
}
public int processWord(String wordString) {
if(wordString.length() < 2) {
return wordString.length();
}
String wordStringES = reduceFromEnd(reduceFromStart(wordString));
if(wordStringES.length() == 1) {
return 1;
}
String wordStringSE = reduceFromStart(reduceFromEnd(wordString));
if(wordString.length() == 1) {
return 1;
}
int aLen;
if(isReduced(wordStringSE)) {
aLen = wordStringSE.length();
} else {
aLen = processWord(wordStringSE);
}
int bLen;
if(isReduced(wordStringES)) {
bLen = wordStringES.length();
} else {
bLen = processWord(wordStringES);
}
return Math.min(aLen, bLen);
}
private boolean isReduced(String wordString) {
int length = wordString.length();
if(length < 2) {
return true;
}
for(int i = 1; i < length; ++i) {
if(wordString.charAt(i) != wordString.charAt(i - 1)) {
return false;
}
}
return wordString.charAt(0) == wordString.charAt(length - 1);
}
private String reduceFromStart(String theWord) {
if(theWord.length() < 2) {
return theWord;
}
StringBuilder buffer = new StringBuilder();
char[] word = theWord.toCharArray();
char curChar = word[0];
for(int i = 1; i < word.length; ++i) {
if(word[i] != curChar) {
curChar = replacementChar(curChar, word[i]);
if(i + 1 == word.length) {
buffer.append(curChar);
break;
}
} else {
buffer.append(curChar);
if(i + 1 == word.length) {
buffer.append(curChar);
}
}
}
return buffer.toString();
}
private String reduceFromEnd(String theString) {
if(theString.length() < 2) {
return theString;
}
StringBuilder buffer = new StringBuilder(theString);
int length = buffer.length();
while(length > 1) {
char a = buffer.charAt(0);
char b = buffer.charAt(length - 1);
if(a != b) {
buffer.deleteCharAt(length - 1);
buffer.deleteCharAt(0);
buffer.append(replacementChar(a, b));
length -= 1;
} else {
break;
}
}
return buffer.toString();
}
public void go() {
Scanner scanner = new Scanner(System.in);
int numEntries = Integer.parseInt(scanner.nextLine());
List<Integer> counts = new LinkedList<Integer>();
for(int i = 0; i < numEntries; ++i) {
counts.add((processWord(scanner.nextLine())));
}
for(Integer count : counts) {
System.out.println(count);
}
}
public static void main(String[] args) {
Sample solution = new Sample();
solution.go();
}
}
This is greedy approach and traversing the path starts with each possible pair and checking the min length.
import java.io.*;
import java.util.*;
class StringSim{
public static void main(String args[]){
Scanner sc = new Scanner(System.in);
StringTokenizer st = new StringTokenizer(sc.nextLine(), " ");
int N = Integer.parseInt(st.nextToken());
String op = "";
for(int i=0;i<N;i++){
String str = sc.nextLine();
op = op + Count(str) + "\n";
}
System.out.println(op);
}
public static int Count( String str){
int min = Integer.MAX_VALUE;
char pre = str.charAt(0);
boolean allSame = true;
//System.out.println("str :" + str);
if(str.length() == 1){
return 1;
}
int count = 1;
for(int i=1;i<str.length();i++){
//System.out.println("pre: -"+ pre +"- char at "+i+" is : -"+ str.charAt(i)+"-");
if(pre != str.charAt(i)){
allSame = false;
char rep = (char)(('a'+'b'+'c')-(pre+str.charAt(i)));
//System.out.println("rep :" + rep);
if(str.length() == 2)
count = 1;
else if(i==1)
count = Count(rep+str.substring(2,str.length()));
else if(i == str.length()-1)
count = Count(str.substring(0,str.length()-2)+rep);
else
count = Count(str.substring(0,i-1)+rep+str.substring(i+1,str.length()));
if(min>count) min=count;
}else if(allSame){
count++;
//System.out.println("count: " + count);
}
pre = str.charAt(i);
}
//System.out.println("min: " + min);
if(allSame) return count;
return min;
}
}
Following NominSim's observations, here is probably an optimal solution that runs in linear time with O(1) space usage. Note that it is only capable of finding the length of the smallest reduction, not the reduced string itself:
def reduce(string):
a = string.count('a')
b = string.count('b')
c = string.count('c')
if ([a,b,c].count(0) >= 2):
return a+b+c
elif (all(v % 2 == 0 for v in [a,b,c]) or all(v % 2 == 1 for v in [a,b,c])):
return 2
else:
return 1
There is some underlying structure that can be used to solve this problem in O(n) time.
The rules given are (most of) the rules defining a mathematical group, in particular the group D_2 also sometimes known as K (for Klein's four group) or V (German for Viergruppe, four group). D_2 is a group with four elements, A, B, C, and 1 (the identity element). One of the realizations of D_2 is the set of symmetries of a rectangular box with three different sides. A, B, and C are 180 degree rotations about each of the axes, and 1 is the identity rotation (no rotation). The group table for D_2 is
|1 A B C
-+-------
1|1 A B C
A|A 1 C B
B|B C 1 A
C|C B A 1
As you can see, the rules correspond to the rules given in the problem, except that the rules involving 1 aren't present in the problem.
Since D_2 is a group, it satisfies a number of rules: closure (the product of any two elements of the group is another element), associativity (meaning (x*y)*z = x*(y*z) for any elements x, y, z; i.e., the order in which strings are reduced doesn't matter), existence of identity (there is an element 1 such that 1*x=x*1=x for any x), and existence of inverse (for any element x, there is an element x^{-1} such that x*x^{-1}=1 and x^{-1}*x=1; in our case, every element is its own inverse).
It's also worth noting that D_2 is commutative, i.e., x*y=y*x for any x,y.
Given any string of elements in D_2, we can reduce to a single element in the group in a greedy fashion. For example, ABCCCCCCC=CCCCCCCC=CCCCCC=CCCC=CC=1. Note that we don't write the element 1 unless it's the only element in the string. Associativity tells us that the order of the operations doesn't matter, e.g., we could have worked from right to left or started in the middle and gotten the same result. Let's try from the right: ABCCCCCCC=ABCCCCC=ABCCC=ABC=AA=1.
The situation of the problem is different because operations involving 1 are not allowed, so we can't just eliminate pairs AA, BB, or CC. However, the situation is not that different. Consider the string ABB. We can't write ABB=A in this case. However, we can eliminate BB in two steps using A: ABB=CB=A. Since order of operation doesn't matter by associativity, we're guaranteed to get the same result. So we can't go straight from ABB to A but we can get the same result by another route.
Such alternate routes are available whenever there are at least two different elements in a string. In particular, in each of ABB, ACC, BAA, BCC, CAA, CBB, AAB, AAC, BBA, BBC, CCA, CCB, we can act as if we have the reduction xx=1 and then drop the 1.
It follows that any string that is not homogeneous (not all the same letter) and has a double-letter substring (AA, BB, or CC) can be reduced by removing the double letter. Strings that contain just two identical letters can't be further reduced (because there is no 1 allowed in the problem), so it seems safe to hypothesize that any non-homogeneous string can be reduced to A, B, C, AA, BB, CC.
We still have to be careful, however, because CCAACC could be turned into CCCC by removing the middle pair AA, but that is not the best we can do: CCAACC=AACC=CC or AA takes us down to a string of length 2.
Another situation we have to be careful of is AABBBB. Here we could eliminate AA to end with BBBB, but it's better to eliminate the middle B's first, then whatever: AABBBB=AABB=AA or BB (both of which are equivalent to 1 in the group, but can't be further reduced in the problem).
There's another interesting situation we could have: AAAABBBB. Blindly eliminating pairs takes us to either AAAA or BBBB, but we could do better: AAAABBBB=AAACBBB=AABBBB=AABB=AA or BB.
The above indicate that eliminating doubles blindly is not necessarily the way to proceed, but nevertheless it was illuminating.
Instead, it seems as if the most important property of a string is non-homogeneity. If the string is homogeneous, stop, there's nothing we can do. Otherwise, identify an operation that preserves the non-homogeneity property if possible. I assert that it is always possible to identify an operation that preserves non-homogeneity if the string is non-homogeneous and of length four or greater.
Proof: if a 4-substring contains two different letters, a third letter can be introduced at a boundary between two different letters, e.g., AABA goes to ACA. Since one or the other of the original letters must be unchanged somewhere within the string, it follows that the result is still non-homogeneous.
Suppose instead we have a 4-substring that has three different elements, say AABC, with the outer two elements different. Then if the middle two elements are different, perform the operation on them; the result is non-homogeneous because the two outermost elements are still different. On the other hand, if the two inner elements are the same, e.g., ABBC, then they have to be different from both outermost elements (otherwise we'd only have two elements in the set of four, not three). In that case, perform either the first or third operation; that leaves either the last two elements different (e.g., ABBC=CBC) or the first two elements different (e.g., ABBC=ABA) so non-homogeneity is preserved.
Finally, consider the case where the first and last elements are the same. Then we have a situation like ABCA. The middle two elements both have to be different from the outer elements, otherwise we'd have only two elements in this case, not three. We can take the first available operation, ABCA=CCA, and non-homogeneity is preserved again.
End of proof.
We have a greedy algorithm to reduce any non-homogeneous string of length 4 or greater: pick the first operation that preserves non-homogeneity; such an operation must exist by the above argument.
We have now reduced to the case where we have a non-homogeneous string of 3 elements. If two are the same, we either have doubles like AAB etc., which we know can be reduced to a single element, or we have two elements with no double like ABA=AC=B which can also be reduced to a single element, or we have three different elements like ABC. There are six permutations, all of which =1 in the group by associativity and commutativity; all of them can be reduced to two elements by any operation; however, they can't possibly be reduced below a homogeneous pair (AA, BB, or CC) since 1 is not allowed in the problem, so we know that's the best we can do in this case.
In summary, if a string is homogeneous, there's nothing we can do; if a string is non-homogeneous and =A in the group, it can be reduced to A in the problem by a greedy algorithm which maintains non-homogeneity at each step; the same if the string =B or =C in the group; finally if a string is non-homogeneous and =1 in the group, it can be reduced by a greedy algorithm which maintains non-homogeneity as long as possible to one of AA, BB or CC. Those are the best we can do by the group properties of the operation.
Program solving the problem:
Now, since we know the possible outcomes, our program can run in O(n) time as follows: if all the letters in the given string are the same, no reduction is possible so just output the length of the string. If the string is non-homogeneous, and is equal to the identity in the group, output the number 2; otherwise output the number 1.
To quickly decide whether an element equals the identity in the group, we use commutativity and associativity as follows: just count the number of A's, B's and C's into the variables a, b, c. Replace a = a mod 2, b = b mod 2, c = c mod 2 because we can eliminate pairs AA, BB, and CC in the group. If none of the resulting a, b, c is equal to 0, we have ABC=1 in the group, so the program should output 2 because a reduction to the identity 1 is not possible. If all three of the resulting a, b, c are equal to 0, we again have the identity (A, B, and C all cancelled themselves out) so we should output 2. Otherwise the string is non-identity and we should output 1.
//C# Coding
using System;
using System.Collections.Generic;
namespace ConsoleApplication1
{
class Program
{
static void Main(string[] args)
{
/*
Keep all the rules in Dictionary object 'rules';
key - find string, value - replace with value
eg: find "AB" , replace with "AA"
*/
Dictionary<string, string> rules = new Dictionary<string, string>();
rules.Add("AB", "AA");
rules.Add("BA", "AA");
rules.Add("CB", "CC");
rules.Add("BC", "CC");
rules.Add("AA", "A");
rules.Add("CC", "C");
// example string
string str = "AABBCCCA";
//output
Console.WriteLine(fnRecurence(rules, str));
Console.Read();
}
//funcation for applying all the rules to the input string value recursivily
static string fnRecurence(Dictionary<string, string> rules,string str)
{
foreach (var rule in rules)
{
if (str.LastIndexOf(rule.Key) >= 0)
{
str = str.Replace(rule.Key, rule.Value);
}
}
if(str.Length >1)
{
int find = 0;
foreach (var rule in rules)
{
if (str.LastIndexOf(rule.Key) >= 0)
{
find = 1;
}
}
if(find == 1)
{
str = fnRecurence(rules, str);
}
else
{
//if not find any exit
find = 0;
str = str;
return str;
}
}
return str;
}
}
}
Here is my C# solution.
public static int StringReduction(string str)
{
if (str.Length == 1)
return 1;
else
{
int prevAns = str.Length;
int newAns = 0;
while (prevAns != newAns)
{
prevAns = newAns;
string ansStr = string.Empty;
int i = 1;
int j = 0;
while (i < str.Length)
{
if (str[i] != str[j])
{
if (str[i] != 'a' && str[j] != 'a')
{
ansStr += 'a';
}
else if (str[i] != 'b' && str[j] != 'b')
{
ansStr += 'b';
}
else if (str[i] != 'c' && str[j] != 'c')
{
ansStr += 'c';
}
i += 2;
j += 2;
}
else
{
ansStr += str[j];
i++;
j++;
}
}
if (j < str.Length)
{
ansStr += str[j];
}
str = ansStr;
newAns = ansStr.Length;
}
return newAns;
}
}
Compare two characters at a time and replace if both adjacent characters are not same. To get optimal solution, run once from start of the string and once from end of the string. Return the minimum value.
Rav solution is :-
int same(char* s){
int i=0;
for(i=0;i<strlen(s)-1;i++){
if(*(s+i) == *(s+i+1))
continue;
else
return 0;
}
return 1;
}
int reduceb(char* s){
int ret = 0,a_sum=0,i=0;
int len = strlen(s);
while(1){
i=len-1;
while(i>0){
if ((*(s+i)) == (*(s+i-1))){
i--;
continue;
} else {
a_sum = (*(s+i)) + (*(s+i-1));
*(s+i-1) = SUM - a_sum;
*(s+i) = '\0';
len--;
}
i--;
}
if(same(s) == 1){
return strlen(s);
}
}
}
int reducef(char* s){
int ret = 0,a_sum=0,i=0;
int len = strlen(s);
while(1){
i=0;
while(i<len-1){
if ((*(s+i)) == (*(s+i+1))){
i++;
continue;
} else {
a_sum = (*(s+i)) + (*(s+i+1));
*(s+i) = SUM - a_sum;
int j=i+1;
for(j=i+1;j<len;j++)
*(s+j) = *(s+j+1);
len--;
}
i++;
}
if(same(s) == 1){
return strlen(s);
}
}
}
int main(){
int n,i=0,f=0,b=0;
scanf("%d",&n);
int a[n];
while(i<n){
char* str = (char*)malloc(101);
scanf("%s",str);
char* strd = strdup(str);
f = reducef(str);
b = reduceb(strd);
if( f > b)
a[i] = b;
else
a[i] = f;
free(str);
free(strd);
i++;
}
for(i=0;i<n;i++)
printf("%d\n",a[i]);
}
#Rav
this code will fail for input "abccaccba".
solution should be only "b"
but this code wont give that. Since i am not getting correct comment place(due to low points or any other reason) so i did it here.
This problem can be solved by greedy approach. Try to find the best position to apply transformation until no transformation exists. The best position is the position with max number of distinct neighbors of the transformed character.
You can solve this using 2 pass.
In the first pass you apply
len = strlen (str) ;
index = 0 ;
flag = 0 ;
/* 1st pass */
for ( i = len-1 ; i > 0 ; i -- ) {
if ( str[i] != str[i-1] ) {
str[i-1] = getChar (str[i], str[i-1]) ;
if (i == 1) {
output1[index++] = str[i-1] ;
flag = 1 ;
break ;
}
}
else output1[index++] = str[i] ;
}
if ( flag == 0 )
output1[index++] = str[i] ;
output1[index] = '\0';
And in the 2nd pass you will apply the same on 'output1' to get the result.
So, One is forward pass another one is backward pass.
int previous = a.charAt(0);
boolean same = true;
int c = 0;
for(int i = 0; i < a.length();++i){
c ^= a.charAt(i)-'a'+1;
if(a.charAt(i) != previous) same = false;
}
if(same) return a.length();
if(c==0) return 2;
else return 1;
import java.util.Scanner;
public class StringReduction {
public static void main(String[] args) {
Scanner sc = new Scanner(System.in);
String str = sc.nextLine();
int length = str.length();
String result = stringReduction(str);
System.out.println(result);
}
private static String stringReduction(String str) {
String result = str.substring(0);
if(str.length()<2){
return str;
}
if(str.length() == 2){
return combine(str.charAt(0),str.charAt(1));
}
for(int i =1;i<str.length();i++){
if(str.charAt(i-1) != str.charAt(i)){
String temp = str.substring(0, i-1) + combine(str.charAt(i-1),str.charAt(i)) + str.substring(i+1, str.length());
String sub = stringReduction(temp);
if(sub.length() < result.length()){
result = sub;
}
}
}
return result;
}
private static String combine(char c1, char c2) {
if(c1 == c2){
return "" + c1 + c2;
}
else{
if(c1 == 'a'){
if(c2 == 'b'){
return "" + 'c';
}
if(c2 == 'c') {
return "" + 'b';
}
}
if(c1 == 'b'){
if(c2 == 'a'){
return "" + 'c';
}
if(c2 == 'c') {
return "" + 'a';
}
}
if(c1 == 'c'){
if(c2 == 'a'){
return "" + 'b';
}
if(c2 == 'b') {
return "" + 'a';
}
}
return null;
}
}
}
JAVASCRIPT SOLUTION:
function StringChallenge(str) {
// code goes here
if(str.length == 1) {
return 1;
} else {
let prevAns = str.length;
let newAns = 0;
while(prevAns != newAns) {
prevAns = newAns;
let ansStr = "";
let i = 1;
let j = 0;
while(i < str.length) {
if(str[i] !== str[j]) {
if(str[i] != 'a' && str[j] != 'a') {
ansStr += 'a';
} else if(str[i] != 'b' && str[j] !='b') {
ansStr +='b';
} else if(str[i] != 'c' && str[j] != 'c') {
ansStr += 'c';
}
i += 2;
j += 2;
} else {
ansStr += str[j];
j++;
i++;
}
}
if(j < str.length) {
ansStr += str[j];
}
str = ansStr;
newAns = ansStr.length;
}
return newAns;
}
}

how do I perform mathematical functions that may exceed LONG_MAX

how do I perform ( A div B ) mod C where I am calculating A in one function,B in another function,and C is say ( 10 pow 9 plus 7 ) but both A and B may be greater than C,or INT_MAX
if the only problem is the size of the numbers you can use long. if the number may be larger then long.MAX_VALUE then you need some function to calculate using strings, or use two or more long types, and make your own functions. for example, the plus function will get two long types, check if the sum of them is smaller then one of them(meaning they went over the MAX_SIZE), then return an array of long, containing to numbers, one for the carry bit, and one for the sum.
here is an example in c#, but it's easy to translate it to c++
public static string sum(long a, long b)
{
string sum;
if (a + b > a && a + b > b)
{
sum = (a + b).ToString();
}
else
{
string aStr = a.ToString();
string bStr = b.ToString();
if (bStr.Length > aStr.Length)
{
string tmp = aStr;
aStr = bStr;
bStr = tmp;
}
sum = new string('0', aStr.Length + bStr.Length);
char[] arr = sum.ToCharArray();
for (int i = 0; i < bStr.Length; i++)
{
int loc = sum.Length - 1 - i;
arr[loc] += (char)(aStr[aStr.Length - 1 - i] + bStr[bStr.Length - 1 - i] - '0' * 2);
if (arr[loc] > '9')
{
arr[loc - 1] = '1';
arr[loc] = (char)(arr[loc] - '9' - 1);
}
}
for (int i = bStr.Length ; i < aStr.Length; i++)
{
int loc = sum.Length - 1 - i;
arr[loc] += (char)(aStr[aStr.Length - 1 - i] - '0');
if (arr[loc] > '9')
{
arr[loc - 1] = '1';
arr[loc] = (char)(arr[loc] - '9' - 1);
}
}
sum = new string(arr);
}
return sum;
}
}

Trimming a string with <= 2 characters

Suppose you are given an input string:
"my name is vikas"
Suggest an algorithm to modify it to:
"name vikas"
Which means remove words having length <=2 or say k characters, to make it generic.
I think you can do this in-place in O(n) time. Iterate over the string, keeping a pointer to begining the word you're processing. If you find that the length of the word is greater than k, you overwrite the begining of the string with this word. Here's a C code (it assumes that each word is separated by exacly on space):
void modify(char *s, int k){
int n = strlen(s);
int j = 0, cnt = 0, r = 0, prev = -1;
s[n++] = ' '; // Setinel to avoid special case
for(int i=0; i<n; i++){
if(s[i] == ' '){
if (cnt > k){
if(r > 0) s[r++] = ' ';
while(j < i) s[r++] = s[j++];
}
cnt = 0;
}
else {
if (prev == ' ') j = i;
cnt++;
}
prev = s[i];
}
s[r] = '\0';
}
int main(){
char s[] = "my name is vikas";
modify(s, 2);
printf("%s\n", s);
}
"a short sentence of words" split ' ' filter {_.length > 2} mkString " "
(Scala)
Iterate over individual characters of String keeping the current position in the string and the "current word", accumulate all current words with length >= k, reassemble String from accumulated words?
This algorithm uses in-place rewriting and minimizes the number of copies between elements:
final int k = 2;
char[] test = " my name is el jenso ".toCharArray();
int l = test.length;
int pos = 0;
int cwPos = 0;
int copyPos = 0;
while (pos < l)
{
if (Character.isWhitespace(test[pos]))
{
int r = pos - cwPos;
if (r - 1 < k)
{
copyPos -= r;
cwPos = ++pos;
}
else
{
cwPos = ++pos;
test[copyPos++] = ' ';
}
}
else
{
test[copyPos++] = test[pos++];
}
}
System.out.println(new String(test, 0, copyPos));
split() by " " and omit if length() <= 2
Something like that will suffice (time complexity is optimal, I guess):
input
.Split(' ')
.Where(s => s.Length > k)
.Aggregate(new StringBuilder(), (sb, s) => sb.Append(s))
.ToString()
What about space complexity? Well, this can run in O(k) (we can't count size of input and output, of course), if you think about it. It won't in .NET, because Split makes actual array. But you can build iterators instead. And if you imagine the string is just iterator over characters, it will become O(1) algorithm.

Resources