我有一个字符串列表(包含室友)如下:
List<string> myList = new List<string> { "11", "11a", "11 a", "11-1", "11a-1" };
myList.Sort(new MyComparer());现在,我希望将列表排序为{ "11", "11a", "11 a", "11a-1", "11-1" },首先是所有没有后缀的软管,然后是后面有字母的,然后是所有其他的。因此,我将字符串拆分为前缀(实际数字)和后缀(数字本身后面的所有内容)。
private class MyComparer : IComparer<string>
{
protected virtual int compareHouseNumbers(string hnr1, string hnr2)
{
// ...
// split number and suffix
// ...
// housenumbers (integers) are also equal so let the suffix decide which one is greater
// the suffixes do not contain any spaces now
if (String.IsNullOrEmpty(suffix1)) return -1;
else if (String.IsNullOrEmpty(suffix2)) return 1;
// the following shell ensure that a letter comes "before" any other special char (such as "-" or "/")
if (Char.IsLetter(suffix1.FirstOrDefault()) && !Char.IsLetter(suffix2.FirstOrDefault())) return -1;
else if (!Char.IsLetter(suffix1.FirstOrDefault()) && Char.IsLetter(suffix2.FirstOrDefault())) return 1;
// if we have more complexity (compare 11a-1 with 11a-2) we use standard string-comparison
return String.Compare(suffix1, suffix2);
}
/// <inheritDoc/>
public int Compare(string lbz1, string lbz2)
{
return this.compareHouseNumbers(lbz1, lbz2);
}
}但我很难根据这些后缀对列表进行排序。我得到的列表是{"11", "11 a", "11a", "11-1", "11a-1"},而交换的条目"11a"和"11 a"是可以的,但就我们的目的而言,我不明白为什么最后一个条目是"11a-1"而不是"11-1"。我已经通过比较进行了调试,但是很明显,这两个成员从来没有被直接比较过,这使得我很难理解这里到底发生了什么。我能做些什么才能让后缀以"a"开头?
如果有一个更优雅的方式来实现这一点,我将开放任何意见,以改进。
编辑:将输入分割成实际的数字和后缀主要是使用这个regex (\\d+)\\s*(\\S*)完成的。这将导致整数部分(实际的大小写)和字符串-在该数字之后的部分。之后,我们只通过使用suffix1 = suffix1.Trim(' ', '-', '/'); (suffix2适当地)从后缀中修剪任何非字母数字字符,在说了这之后,我们实际上并不将-1与a-1进行比较,而是将1与a-1进行比较。但是,这不会改变结果本身的任何内容(因为-1和1在选择上都比a小)。
EDIT2:我删除了列表中的一些成员,以便只剩下两个有问题的成员:List<string> myList = new List<string> { "11-1", "11a-1" };已经改变了这一点,排序的结果和预期的一样:{ "11a-1", "11-1" }
EDIT3:我刚刚更改了列表中成员的顺序(将11放在列表的末尾)。现在的结果也和预期的一样。因此,它似乎取决于列表中元素的初始顺序。很奇怪..。
发布于 2015-01-14 12:30:32
尝试如下,您没有给出如何分割前缀和后缀的示例,因此我编写了一个示例:
private class MyComparer : IComparer<string>
{
private static readonly Regex matchRegex = new Regex(@"^(?<prefix>\d+)(?<spaces>\s*)(?<suffix>.*?)$");
private int match(string compare, out string prefix, out string suffix) {
var match= matchRegex.Match(compare);
prefix=match.Groups["prefix"].Value;
suffix=match.Groups["suffix"].Value;
return match.Groups["spaces"].Value.Length;
}
protected virtual int compareHouseNumbers(string hnr1, string hnr2)
{
// ...
// split number and suffix
// ...
string prefix1;
string prefix2;
string suffix1;
string suffix2;
var spaces1 = match(hnr1, out prefix1,out suffix1);
var spaces2 = match(hnr2, out prefix2,out suffix2);
Debug.WriteLine("Comparing '{0}' and '{1}'",suffix1,suffix2);
var prefixCompare = String.Compare(prefix1,prefix2);
if (prefixCompare != 0) {
return prefixCompare;
}
// housenumbers (integers) are also equal so let the suffix decide which one is greater
// the suffixes do not contain any spaces now
// FIX IS HERE!!!
// Previous code would compare "11" and "11" and return -1 which confuses the sort
if (String.IsNullOrEmpty(suffix1)) return (String.IsNullOrEmpty(suffix2)) ? 0 : -1;
else if (String.IsNullOrEmpty(suffix2)) return 1;
// the following shell ensure that a letter comes "before" any other special char (such as "-" or "/")
if (Char.IsLetter(suffix1.FirstOrDefault()) && !Char.IsLetter(suffix2.FirstOrDefault())) return -1;
else if (!Char.IsLetter(suffix1.FirstOrDefault()) && Char.IsLetter(suffix2.FirstOrDefault())) return 1;
// if we have more complexity (compare 11a-1 with 11a-2) we use standard string-comparison
var result = String.Compare(suffix1, suffix2);
// if the suffixes are equal sort on the number of spaces between prefix and suffix
if (result == 0) {
return (spaces1 - spaces2) <0 ? -1 : (spaces1 == spaces2) ? 0: 1;
}
return result;
}
/// <inheritDoc/>
public int Compare(string lbz1, string lbz2)
{
return this.compareHouseNumbers(lbz1, lbz2);
}
}https://stackoverflow.com/questions/27942444
复制相似问题