"This is the best moth".match(/th/gi);
or with a variable for your string
var string = "This is the best moth";
alert(string.match(/th/gi));
in a regex is a word boundary so th
will only match a th
that at the beginning of a word.
gi
is for a global match (look for all occurrences) and case insensitive
(I threw moth
in there to as a reminder to check that it is not matched)
jsFiddle example
Edit:
So, the above only returns the part that you match (th
). If you want to return the entire words, you have to match the entire word.
This is where things get tricky fast. First with no HTML entity letter:
string.match(/th[^]*?/gi);
Example
To match the entire word go from the word boundary
grab the th
followed by non word boundaries [^]
until you get to another word boundary
. The *
means you want to look for 0 or more of the previous (non word boundaries) the ?
mark means that this is a lazy match. In other words it doesn't expand to as big as would be possible, but stops at the first opportunity.
If you have HTML entity characters like ä (ä
) things get complicated really fast, and you have to use whitespace or whitespace and a set of defined characters that may be at word boundaries.
string.match(/sth[^s]*|^th[^s]*/gi);
Example with HTML entities.
Since we're not using word boundaries, we have to take care of the beginning of the string separately (|^
).
The above will capture the white space at the beginning of words. Using
will not capture white space, since
has no width.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…