number multiply by 100 when using bulk insert

[PichaiSBG]

I use msnodesqlv8 v.4.5.0, Nodejs v. 21.5.0, MSSQL 2022 in docker to test bulk insert. original data have 14 rows. there are 2 rows that incorrectly. Even I reduce number of row and column to be 1 as I provide below. the inconsistency still there.
Please advice. Thanks.

following are code, json data and result
TABLE structure : CREATE TABLE [dbo].[Z1](
[ABC] [numeric](12, 2) NULL,
)

JSON DATA : [{"ABC":16763.60}]

result after do bulk insert:
select * from Z1
ABC: 1676359.99

const sql = require("msnodesqlv8");
const fs = require("fs");

main();
async function main () {
const driver = "ODBC Driver 17 for SQL Server";
const connectionString = "Driver={" + driver + "};Server=" + server + ";" +
(useTrustedConnection == 'true' ? "Trusted_Connection={Yes};" : "UID=" + user + ";PWD=" + pwd + ";") + "Database={" + database + "};";
const parsedJSON = JSON.parse(fs.readFileSync('data.json','utf8'));
const tableName='Z1';
let conn; // Declare connection outside to ensure it's closed in finally
try {

    conn = await new Promise((resolve, reject) => {
        sql.open(connectionString, function (err, connection) {
            if (err) {
                console.error("Connection Error:", err);
                return reject(err);
            }
            console.log('Successfully connected to SQL Server.');
            resolve(connection);
        });
    });

    // Crucial for correct handling of dates/times based on local server time
    // If your database stores UTC and you want to insert UTC, you can remove this or set to true.
    conn.setUseUTC(false); 
    console.log(`Attempting to bind to table: ${tableName}`);

    const tm = conn.tableMgr();

    // Use a Promise wrapper for tm.bind for better async flow
    const bulkMgr = await new Promise((resolve, reject) => {
        tm.getTable(tableName, (bindErr, manager) => {
            if (bindErr) {
                console.error("Table Binding Error:", bindErr);
                return reject(bindErr);
            }
            if (!manager) {
                // This can happen if tableMgr.bind doesn't return a manager,
                // indicating the table was not found or accessible.
                return reject(new Error(`Could not bind to table '${tableName}'. Ensure it exists and is accessible.`));
            }
            console.log(`Successfully bound to table: ${tableName}. Starting bulk insert.`);
            resolve(manager);
        });
    });

    if (parsedJSON.length === 0) {
        console.log("No data to insert. Skipping bulk insert.");
        return;
    }

    // Perform the bulk insert
    // Use a Promise wrapper for bulkMgr.insertRows for better async flow
    await new Promise((resolve, reject) => {
        bulkMgr.insertRows(parsedJSON, (insertErr, result) => {
            if (insertErr) {
                console.error("Bulk Insert Error:");
                return reject(insertErr);
            }
            // The 'result' object might contain information like rows affected,
            // depending on the driver version and operation.
            console.log(`Bulk insert successful! Result:`, result);
            resolve(result);
        });
    });

} catch (error) {
    // Consolidated error handling
    console.error("An error occurred during the bulk insert process:");
    // You can add more specific error handling here based on 'error' object properties if needed
    // For example:
    // if (error.code === 'ELOGIN') { ... }
    // if (error.message.includes('Invalid column name')) { ... }
} finally {
    // Ensure the connection is always closed
    if (conn) {
        conn.close(() => {
            console.log('Database connection closed.');
        });
    }
}

}

[PichaiSBG]

I have tested by preparing object array that have 10 columns (A-J). first row has number start from 0.00 to 0.09 increase by 0.1 next row until 9999.99 at last column of last row. I found that that are a lot of inconsistencies in original number and its result after the bulk insert.