Numerical, amount class precision loss problem

Problem Description

• In daily projects, the amount of money involves numerical calculation, and the problem of loss of numerical precision after addition, subtraction, multiplication and division often occurs.

Analyze possible causes

• The binary representation of decimal numbers can be imprecise.
• Type mismatch between the numbers used (for example, mixing float and double).

solve

//获取小数位数
decNum(a){
   let r = 0
    a = a.toString()
    if(a.indexOf(".") !== -1){
        r = a.split(".")[1].length
    }
    return r
},
//type:加减乘除
calc(a,b,type){
    let r = 0
    let da = this.decNum(a)
    let db = this.decNum(b)
    let dsum = da + db
    let dmin = Math.min(da,db)
    let dmax = Math.max(da,db)
    dsum += dmax - dmin
    dsum = Math.pow(10,dsum)
    dmax = Math.pow(10,dmax)
    a = parseInt(a,toString().replace(".",""))
    b = parseInt(b,toString().replace(".",""))
    if(da > db){
        b *= Math.pow(10,da-db)
    }else {
        a *= Math.pow(10,db-da)
    }
    switch(type){
        case "add":
            r = (a + b) / dmax
            break
         case "subtract":
            r = (a - b) / dmax
            break
         case "multiply":
            r = (a * b) / dsum
            break
         case "divide":
            r = a / b
            break
    }
    return r
}

Common applications

• In the project, the floating-point value type of the amount involved will be used