如何拆分 java 中组件的数学表达式？

Question

I have some issue about split mathematical expression.

I want to split every number, operator and parentheses, and send them to an array, this is my code for different case:

String infix = "-(1/10)^(-2)-(-2)^(5)-2-3+1^(-2)";

//List<String>    tokenList = Arrays.asList(infix.split("(?<=[+*/()!^-])|(?=[+*/()!^-])"));
//List<String>    tokenList = Arrays.asList(infix.split("(?=[-+*/()])|(?<=[^-+*/][-+*/])|(?<=[()])"));
//List<String>     tokenList = Arrays.asList(infix.split("((?<=\\^)|(?=\\^))"));
List<String>     tokenList = Arrays.asList(infix.split("((?<=[+*/()!])|(?=[+*/()!]))|((?<=\\^)|(?=\\^))|([0-9]+(?<=[-])|(?=[-]))"));

i can't separate it. I often got wrong expression

[1, 10, /, -2, ^, -2, 5, -2, -3, ^, 1, -2, ^, +, -, -]

My program can't divide minus sign. if have -1-1 my program divide it like [-1,-1]. Please, help me solve this issue.

Answer 1

Agree with Aaron. The solution could be based on FSM (finite state machine). Essentially, you have a string expression, and you are starting parsing. Also, consider there's a temporary string variable that is empty for now. When you read the first char, you go and write it to the variable and decide, what it is: in this example, it could be a number, an operator, or a space. You jump to a special state (which should also be stored somewhere). Then you read the next char in the expression, and now you have the history (the current state and the previously read chars). Here you can decide if the token is ended or not, and depending on that you can jump to another state.

For example, you read "1" . You know, that, for instance, it could be a number of any length. You persist "1" in a variable and read the next char. If there's another digit, you could be sure that there's still the same number. If there's a space, or another operator, you push the "1" from the first step to a list of tokens or wherever, clear the string, and persist the next character in it.

Quite often it could be implemented using a switch-case operator.

Answer 2

I found answers on my issue ( code below). It's very easier way to solve this problem.

List<String>     tokenList = Arrays.asList(infix.split(
        "((?<=\\^)|(?=\\^))|((?<=\\s[-+]\\s)" +
        "|(?=\\s[-+]\\s))|((?<=[()])|(?=[()]))" +
        "|((?<=[*/])|(?=[*/]))" +
        "|((?<=\\s[*/]\\s))"));