从IIS获取ERR_CONNECTION_RESET，日志中没有信息(Getting ERR_CONNECTION_RESET from IIS with no information in logs)

 “最佳”取决于您拥有的资源。 
 
 这是我要做的： 
 
 
 
 为每个单词长度存储单独的字典。 
 
 
 我假设您的搜索不区分大小写！ 如果字符a..z出现在字中，则为每个字存储32位值，其中设置位0到25。 将单词存储在Map<Integer, List<String>> （键是32位值，value是此键的所有单词的列表） 
 
 
 如何搜索： 
 
 
 
 拿着想要长度的字典 
 
 
 为您的模式创建32位值。 
 
 
 迭代Map所有键，并检查键的按位和模式的32位值是否等于模式的32位值。 如果不是这不能匹配。 如果通过此检查，则仅仅匹配是不够的，因为它不处理字符的顺序或者如果出现多于1的字符。 但检查速度非常快，不需要查看单词的每个字符。 
 
 
 迭代Map的列表，并通过将列表中每个单词的字符与模式进行比较来检查它们与模式的真实匹配。 
 
 
 示例：三字母单词的字典：the，cat，bag，nor，ega，atc，ron; 
 
-> Hashvalues
     00000000000010000000000010010000   the
     00000000000010000000000000000101   cat, atc
     00000000000000000000000001000011   bag
     00000000000000100110000000000000   nor, ron
     00000000000000000000000001010001   age, ega

 Value for pattern _a_ is 00000000000000000000000000000001

Step 3 returns that the keys 
      00000000000010000000000000000101,  
      00000000000000000000000001000011 and 
      00000000000000000000000001010001 are candidates for matches.



Step 4 returns: 'cat' and 'bag'

'Best' depends on the resources you have. 
 
Here is what I would do:
 
 
 
Store a separate dictionary for each word length.
 
 
I assume your search is not case sensitive! Store for each word a 32-bit value in which the bits 0 to 25 are set if the character a..z occurs in the word. Store the words in a Map<Integer, List<String>> (key is the 32-bit value, value is a list of all words for this key)
 
 
How to search:
 
 
 
Take the dictionary with the words of the wanted length
 
 
Create the 32-bit value for your pattern. 
 
 
Iterate all keys of the Map and check if the key bitwise and with the 32-bit value for your pattern is equal to the 32-bit value for your pattern. If not this cannot be matches. If this check is passed, it is not sufficient to be a match, because it does not handle order of characters or if the occur more than ones. But the check is very fast and does not need to look at each character of the words.
 
 
Iterate the list in the Map and check them which are real matches for your pattern by comparing the characters of each word in the list with your pattern.
 
 
Example: Dictionary for 3-letter words: the, cat, bag, nor, ega, atc, ron;
 
-> Hashvalues
     00000000000010000000000010010000   the
     00000000000010000000000000000101   cat, atc
     00000000000000000000000001000011   bag
     00000000000000100110000000000000   nor, ron
     00000000000000000000000001010001   age, ega

 Value for pattern _a_ is 00000000000000000000000000000001

Step 3 returns that the keys 
      00000000000010000000000000000101,  
      00000000000000000000000001000011 and 
      00000000000000000000000001010001 are candidates for matches.



Step 4 returns: 'cat' and 'bag'