T-sql (sql server 18)

第一页的解决方案副本

CREATE FUNCTION dbo.getExcelColumnNameByOrdinal(@RowNum int)  
RETURNS varchar(5)   
AS   
BEGIN  
    DECLARE @dividend int = @RowNum;
    DECLARE @columnName varchar(max) = '';
    DECLARE @modulo int;

    WHILE (@dividend > 0)
    BEGIN  
        SELECT @modulo = ((@dividend - 1) % 26);
        SELECT @columnName = CHAR((65 + @modulo)) + @columnName;
        SELECT @dividend = CAST(((@dividend - @modulo) / 26) as int);
    END
    RETURN 
       @columnName;

END;

我在我的第一篇文章中发现了一个错误，所以我决定坐下来算算。我发现用来识别Excel列的数字系统不是另一个人说的26进制系统。以10为基数考虑以下情况。你也可以用字母表中的字母来做这件事。

空间 :.........................S1, s2, s3: S1, s2, s3 ....................................0,00, 000:..A aa aaa ....................................1,01, 001:..B ab aab ....................................…，…，…:……，…，… ....................................9,99,999:..Z, zz, ZZZ 空间中的总状态:10,100,1000:26,676,17576 国家总 :............... 1110年 ................ 18278年

Excel在以26为基数的字母空格中对列进行编号。你可以看到，一般来说，状态空间的级数是a, a^2, a^3，…对于以a为底的情况，状态的总数是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.

使用基本索引为0的c#代码是

    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;
    }

/ /旧邮政

c#中的零基础解决方案。

    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;
    }

private String getColumn(int c) {
    String s = "";
    do {
        s = (char)('A' + (c % 26)) + s;
        c /= 26;
    } while (c-- > 0);
    return s;
}

它不是以26为底，系统中没有0。如果有的话，'Z'后面应该是'BA'而不是'AA'。

这些我的代码转换特定的数字(索引从1开始)到Excel列。

    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;
    }

我的单元测试:

    [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));
    }

递归很简单。

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) );
}

这是所有其他人以及谷歌重定向到的问题，所以我在这里发布这个。

这些答案中有许多是正确的，但对于简单的情况来说太麻烦了，比如当您的列不超过26个时。如果你怀疑你是否会进入双字符列，那么忽略这个答案，但如果你确定你不会，那么你可以在c#中简单地这样做:

public static char ColIndexToLetter(short index)
{
    if (index < 0 || index > 25) throw new ArgumentException("Index must be between 0 and 25.");
    return (char)('A' + index);
}

见鬼，如果你对你传递的东西有信心，你甚至可以删除验证并使用内联:

(char)('A' + index)

这在许多语言中是非常相似的，因此您可以根据需要进行调整。

同样，只有在100%确定不超过26列的情况下才使用这种方法。