Javascript split the string

"(length>10)&(length<100)".split( /([()><&])/ ).filter( Boolean )
["(", "length", ">", "10", ")", "&", "(", "length", "<", "100", ")"]

This splits at either: (, ), >, < or & (the "or" is thanks to the [] around).

Keeping the split characters is done thanks to the capture (the parentheses around the square brackets - it's ES5 though, so not supported in IE8 and below).

Finally, to remove empty strings, I use filter( Boolean ) on the array ( ES5 too, not supported in IE8 and below either).

result = subject.split(/\b|(?!\w)/);

This splits at boundaries between alphanumeric and non-alphanumeric characters, additionally between two non-alnum characters. You might get an empty match at the start/end of the string, so you need to discard zero-length results.

Instead of a split, I'd go for a global match, which behaves more like a tokenizer:

var input = "(length>10)&(length<100)";
var tokens = input.match(/\d+|[a-zA-Z]\w*|[()]|[<>=&|]+/g);

It scans the input and matches the following patterns (in order):

\d+ # one ore more digits
| # OR
[a-zA-Z]\w* # an identifier
| # OR
[()] # a single opening- or closing parenthesis
| # OR
[<>=&|]+ # one or more operators: '<=', '&', '|=', ...

• javascript
• regex