How to decode a barcode into ISBN? - algorithm

For this question with illustrative purpose I will write Javascript code, but that's just an illustration, the question is language-agnostic.
I need to write a function which takes a barcode text (not the image) as an input and returns the ISBN as output. ISBN can be of 10 digits (older books) or 13 digits (newer books). We also know that the last digit of an ISBN is a checksum, which is computed differently if the ISBN is 10 digits long and differently if the ISBN is 13 digits long.
Assuming that input is a string, we can validate whether it is a valid ISBN, like:
function isValidISBN10(input) {
if (input.length !== 10) return false;
var sum = 0;
var p = 10;
for (var index = 0; index < 10; index++) {
sum += ((input[index] === 'X') ? 10 : input[index]) * (p--);
}
return sum % 11 === 0;
}
and ISBN13 can be validated like:
function isValidISBN13(input) {
if (input.length !== 13) return false;
var sum = 0;
var p = 3;
for (var index = 0; index < 13; index++) {
sum += input[index] * (p = (p + 2) % 4);
}
return sum % 10 === 0;
}
Checking a valid ISBN is:
function isValidISBN(input) {
return isValidISBN10(input) || isValidISBN13(input);
}
As we can see, the last digit of an ISBN is the number we should add in order to make sure the result is divisible by 11 (in the case of ISBN10) and 10 (in the case of ISBN13). The 'X' in the case of ISBN10 represents a number of 10 in 11-base.
As far as I understand these articles:
https://www.barcodefaq.com/1d/isbn/
https://isbn-information.com/isbn-barcode.html
barcodes will contain the digits of ISBNs, except its last digit, the example the first article gives is
ISBN = 09767736X
Barcode = 9780976773665
What confuses me is the number of 51050 on this picture
I wonder whether it is part of the barcode or not. If we consider it not to be a barcode, then converting a barcode to an ISBN would be trivial:
function convertBarcodeIntoISBN(input) {
var isbn = {isbn13: input};
if (input.startsWith("978")) {
var isbn10 = input.substring(3);
var checksum = 0;
var p = 10;
for (var index = 0; index < 9; index++) {
checksum += isbn10[index] * (p--);
}
checksum = 11 - (checksum % 11);
if (checksum === 10) checksum = 'X';
isbn10 += checksum;
isbn.isbn10 = isbn10;
}
return isbn;
}
But if we consider 51050 to be part of the barcode, then we will need to mine the ISBN from the barcode, however, in this case I am not sure how should I operate. The best I can pull from the top of my mind is:
function getLastISBNDigit(input) {
if ((input.length != 10) && (input.length != 13)) return;
var is10 = (input.length === 10);
var sum = 0;
var p = (is10 ? 11 : 3);
for (var index = 0; index < input.length - 1; index++) {
sum += ((input[index] === 'X') ? 10 : input[index]) * (p = (is10 ? (p - 1) : ((p + 2) % 4)));
}
var moduloClass = (is10 ? 11 : 10);
var result = (moduloClass - (sum % moduloClass)) % moduloClass;
return ((result === 10) ? 'X' : result);
}
function getISBN(input) {
var isbn = {};
if (input.length > 13) return getISBN(input.substring(0, 13));
if (input.length === 10) {
if (isValidISBN(input)) {
isbn.isbn10 = input;
isbn.isbn13 = "978" + input;
isbn.isbn13 = isbn.isbn13.substring(0, 12) + getLastISBNDigit(isbn.isbn13);
}
} else if (input.length === 13) {
if (isValidISBN(input)) {
isbn.isbn13 = input;
if (input.startsWith("978")) {
isbn.isbn10 = input.substring(3);
isbn.isbn10 = isbn.isbn10.substring(0, 9) + getLastISBNDigit(isbn.isbn10);
}
} else if (input.startsWith("978")) {
return getISBN(input.substring(3));
}
}
return isbn;
}
This is how I think barcodes should be converted into ISBN and ISBN13 values. Am I right with my reasoning?

The second part is the human-readable price (from this slide):
Hence, the first part of your consideration makes sense and 51050 is not part of the barcode!
The price of the product is 10.50$.

Related

why not pass all test case in problem balanced numbers in codewars?

Currently i am doing a challenge in codewars on Balanced numbers and i wrote the code in dart and it successfully completed 100 test cases but for long numbers it is not working properly...So i think that it need some conditions for this long numbers:
String balancedNum(numb) {
// your code here
var numb1 = numb.toString();
List a = numb1.split("");
var left_sum = 0;
var right_sum = 0;
for (var i = 0; i <= a.length / 2; i++) {
left_sum += int.parse(a[i]);
}
List<String> b = a.reversed.toList();
//print(b);
for (var i = 0; i < b.length / 2; i++) {
right_sum += int.parse(b[i]);
}
//print(right_sum);
if (left_sum == right_sum) {
return 'Balanced';
} else {
return 'Not Balanced';
}
}
link to the challenge:https://www.codewars.com/kata/balanced-number-special-numbers-series-number-1/train/dart
The compiler verdict for the code is wrong answer because of a simple logical error.
This is because the length of the list has not been calculated correctly
Kindly have a look at the corrected code below, which passes all the 110 test on the platform:
String balancedNum(numb) {
// your code here
var numb1 = numb.toString();
List a = numb1.split("");
var left_sum = 0;
var right_sum = 0;
double d = ((a.length-1) / 2);
int len = d.floor();
print(len);
for (var i = 0; i < len; i++) {
left_sum += int.parse(a[i]);
}
List<String> b = a.reversed.toList();
print(b);
for (var i = 0; i < len; i++) {
right_sum += int.parse(b[i]);
}
print(left_sum);
print(right_sum);
if (left_sum == right_sum) {
return 'Balanced';
} else {
return 'Not Balanced';
}
}
Your problem is not about "long numbers" since you code also fails for the test case using the number 13.
You properly did not read the following rules:
If the number has an odd number of digits then there is only one middle digit, e.g. 92645 has middle digit 6; otherwise, there are two middle digits , e.g. 1301 has middle digits 3 and 0
The middle digit(s) should not be considered when determining whether a number is balanced or not, e.g 413023 is a balanced number because the left sum and right sum are both 5.
So you need to test if the number are even or uneven since we need to use that information to know how many digits we should use in the sum on each side.
I have done that in the following implementation:
String balancedNum(int numb) {
final numb1 = numb.toString();
final a = numb1.split("");
var split = (a.length / 2).floor();
if (a.length % 2 == 0) {
split--; // the number is even
}
var left_sum = 0;
var right_sum = 0;
for (var i = 0; i < split; i++) {
print(a[i]);
left_sum += int.parse(a[i]);
}
final b = a.reversed.toList();
for (var i = 0; i < split; i++) {
right_sum += int.parse(b[i]);
}
if (left_sum == right_sum) {
return 'Balanced';
} else {
return 'Not Balanced';
}
}

Question about fromCharCode and ending an "if" statement

I'm working on a simple cipher algorithm and I'm trying to figure out how to stop the if statement at character 122 (z) and start back at char 97 (a) once the character count exceeds 122. This is what I have so far and have looked around on MDN and W3 schools and haven't come up with anything.
enter code here
function simpleCipher(str) {
var newString = [];
for (var i = 0; i < str.length; i ++) {
if (str.charCodeAt(i) > 97 || str.charCodeAt(i) < 122) {
var convertString = String.fromCharCode(str.charCodeAt(i) + 4);
var powerString = newString.push(convertString);
} else {
return;
}
}
return newString.join('');
}
Use modular arithmetic.
var aPos = 97;
var zPos = 122;
var charsCount = zPos - aPos + 1;
...
if (aPos <= str.charCodeAt(i) && str.charCodeAt(i) <= zPos) { // probably you want && here
var charNumber = str.charCodeAt(i) - aPos;
var charNumberEncoded = (charNumber + 4) % charsCount;
var convertString = String.fromCharCode(aPos + charNumberEncoded);
var powerString = newString.push(convertString);
}
It is considered a good practice to give constants names and not use numbers in the code. Such numbers often referenced as magic numbers and make it harder to read the code.

Czech national identity card validation

I'm looking for CZ national idenity card number validation algorithm. There can be 9 or 10 digits sequenced by special algoritm, but there are not just numbers :) I've found this function:
function checkId(id) {
var x, sum_nb=0, check;
var steps = new Array(7,3,1,7,3,1,7,3);
id = id.toUpperCase(),
splitted = id.split("");
if (splitted.length != 9 && splitted.length != 10) {return false;}
for (x = 0; x < 9; x++)
{
if (x == 0 || x == 1 || x == 2) {
sum_nb += steps[x] * parseInt(id.charCodeAt(x)-55);
}
if (x > 3) {
sum_nb += steps[x-1] * parseInt(id.charAt(x));
}
}
check = sum_nb % 10;
if (check == id.charAt(3))
return true;
else
return false;
}
But I'm not sure that it's correct. Is anyone knows correct algoritm? It can be written on any language, it does not matter.

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;
}

Making your own wildcard and recognize patterns to zip your list

I have an algorithm trouble, I don't know how to do it and how to call it (Does it have any specific name?)
For example, if I have this sequence:
000 CASE_01
001 CASE_02
010 CASE_03
011 CASE_02
100 CASE_02
101 CASE_01
110 CASE_01
111 CASE_01
I want to convert it in something like this:
000 CASE_01
0-1 CASE_02
010 CASE_03
100 CASE_02
1-1 CASE_01
11- CASE_01
I have called it wildcard because I think is the most correct way...
They not necessarily have 3 bits, you must do it with n bits
If only I had the psudo-code I could write it to any language (Python in my way)
The code below generates all possible wildcard strings for the number of bits encountered in the input (e.g. --, -0, -1, 0-, 00, 01, 1-, 10 and 11 for 2 bits), and also creates patterns where numbers are inverted and wildcards become 1 (so wildcard 0-1 has pattern 110), and masks where numbers become 1 and wildcards become 0 (so wildcard 0-1 has mask 101).
The binary numbers in the input are then XOR-ed with the patterns and AND-ed with the masks to check whether they fit a certain wildcard. If a wildcard has the required number of matching numbers (2 ^ number_of_wildcards), it is added to the output and the matching numbers are removed from the input.
Run the code snippet to see the algorithm in action with your example input, to which I added a fourth case with larger binary numbers.
function wildcard(input) {
var output = [], cases = [], wilds = [], patts = [], masks = [];
var bits = groupCases(cases);
for (var i = 0; i <= bits; i++) wilds[i] = [];
wildStrings(bits);
convertStrings(wilds, patts, "-01", "110");
convertStrings(wilds, masks, "-01", "011");
for (var c = 0; c < cases.length; c++) {
for (var i = 0, j = Math.pow(2, bits); i <= bits; i++, j /= 2) {
for (var k = 0; k < patts[i].length; k++) {
var patt = patts[i][k];
var mask = masks[i][k];
var matches = [];
for (var d = 0; d < cases[c].nums.length; d++) {
var num = cases[c].nums[d];
if (((num ^ patt) & mask) == mask) matches.push(d);
}
if (matches.length == j) {
output.push(wilds[i][k] + " " + cases[c].id);
for (var l = j - 1; l >= 0; l--) cases[c].nums.splice(matches[l], 1);
}
}
}
}
return output;
function groupCases(cases) {
var max = 0;
for (var i = 0; i < input.length; i++) {
var num = parseInt(input[i], 2);
if (num > max) max = num;
var id = input[i].slice(input[i].indexOf(" ") + 1);
var pos = 0;
while (cases[pos] != undefined && cases[pos].id != id) ++pos;
if (cases[pos] == undefined) cases[pos] = {id: id, nums: []};
cases[pos].nums.push(num);
}
return Math.ceil(Math.log(max) / Math.log(2));
}
function wildStrings(len, wild, str) {
wild = wild || 0;
str = str || "";
for (var i = 0; i < 3; i++) {
var w = (i == 0) ? 1 : 0;
var s = str + ["-","0","1"][i];
if (len > 1) wildStrings(len - 1, wild + w, s)
else wilds[bits - wild - w].push(s);
}
}
function convertStrings(input, output, from, to) {
for (var i = 0; i < input.length; i++) {
output[i] = [];
for (var j = 0; j < input[i].length; j++) {
var str = input[i][j], s = "";
for (var k = 0; k < str.length; k++) {
s += to.charAt(from.indexOf(str.charAt(k)));
}
output[i].push(parseInt(s, 2));
}
}
}
}
var a = ["000 CASE_01", "001 CASE_02", "010 CASE_03", "1010 CASE_04",
"011 CASE_02", "100 CASE_02", "1110 CASE_04", "101 CASE_01",
"110 CASE_01", "1100 CASE_04", "1000 CASE_04", "111 CASE_01"];
var w = wildcard(a);
document.write(w.length + " wildcards:<BR><PRE>");
for (var i in w) document.write(w[i] + "<BR>");

Resources