Avoiding scientific notation

Question

I am experimenting using Go to interact with a Database and am running into issues when dealing with Decimal fields.

In this database most fields that are basically integer fields are typed as decimal with 0 precision... for example:

• Date fields are stored in YYYYMMDD format as decimal(8, 0)
• Id numbers are stored as decimal(9, 0)

Basically any int is stored as a decimal with 0 precision.

What I am attempting is to fill fields in a struct....

type Record struct {
    ID          uint
    CHANGE_DATE uint
    ...
}

But when I get the fields from the database, they often come back in a format like this but only if the number is long enough: 2.0141208e+07

I have found that I can Scan into a float and then convert the float to an uint like this..

mydate := float32(0)

for rows.Next() {
    r := Record{}
    row.Scan(&.ID, &mydate)
    myrecord.Change_date = uint(mydate)
}

If the ID is a large enough number then ParseInt fails and I find I have to do the float/int conversion as shown above.

Since I have many fields (almost all numbers) that I would need to do this with, I am wondering if there is a better way to go about this type conversion?

It may also be worth mentioning that this is a database from a packaged ERP system, so changing the table definitions is not an option and the reason I don't just make all fields float in Go is that I am trying to output json and I get the scientific notation in the json output.

Answer 1

I encountered the same issue when parsing JSON from a RPC server. I created a function to convert the scientific notation string back to uint like this:

func scientificNotationToUInt(scientificNotation string) (uint, error) {
    flt, _, err := big.ParseFloat(scientificNotation, 10, 0, big.ToNearestEven)
    if err != nil {
        return 0, err
    }
    fltVal := fmt.Sprintf("%.0f", flt)
    intVal, err := strconv.ParseInt(fltVal, 10, 64)
    if err != nil {
        return 0, err
    }
    return uint(intVal), nil
}

Playground example