问题是您只获得包含\b-单词边界的匹配项.由于星号是非单词字符，因此将其从匹配项中删除，因此解决方案是允许单词边界或星号(\*|\b):

The problem was that you were only getting matches that include a \b - a word boundary. Since an asterisk is a non-word character, it was eliminating it from the match, so the solution was to allow for either a word boundary or an asterisk (\*|\b):

preg_replace('/([a-z0-9.]+)((\*|\b)(?<!or|and|not))/i', '$0'."[45]", $term);

但是，使用负前瞻进行操作会更简单:

However, it's simpler to do it with a negative lookahead:

preg_replace('/\b(?!or|and|not)([a-z0-9*.]+)/i', '$0'."[45]", $term);

注意:在字符类中，星号和点号不是元字符，因此不需要像原始表达式中那样转义它们:[a-z0-9\*\.]+.

Note: Within character classes asterisks and periods are not metacharacters, so they don't need to be escaped as you had in your original expression: [a-z0-9\*\.]+.