用于验证单词的正则表达式不包含数字或特殊字符

Question

I am developing a java app, running on android. I am trying to pick all words which do not contain any embedded digits or symbols.

The best I have come up with is:

\b[a-zA-Z]+[a-zA-Z]*+\b

Test Data:

this is a test , an0ther gr8 WW##ee one, w1n 1test test1 end

This results in picking the following: this, is, a, test, WW##ee, one, end

I need to eliminate the WW##ee from the results.

Answer 1

You shouldn't use a word boundary meta-character \\b since it matches the position right after WW which sees a hash # character. This position is a word boundary itself. So you should pick up a different way:

(?<![\S&&[^,]])[a-zA-Z]+(?![\S&&[^,]])

Using character class intersection feature of Java's regex you are able to define punctuation characters that are allowed to follow or precede a word character. Here it is a comma , .

Answer 2

您可以使用“向后看”和“向前看”来检查是否没有# 。

\b(?<!\#)[a-zA-Z]+(?!\#)\b

Answer 3

My solution has evolved a bit as I have gotten additional help with this. So, this is now my best solution but still a bit lacking. I have not been able to accept "as-is" while rejecting "-this-" and a similar case of accept "and/or" while rejecting "/slash/". Also for simplicity I have made the input data single word per line.

^(?:[\\p{P}\\p{S}]) ?((?:[\\p{L}\\p{Pd}'])+)(?:[\\p{P}\\p{S}]) $

as-is is picked valid

-this- is valid but I wish it weren't

and/or is not valid but I wish it would be picked

/slash/ "slash" is picked valid

(test) "test" is picked valid

[test] "test" is picked valid

<test> "test" is picked valid