Euler to quaternion conversion error

edit

I try to convert Euler values into quaternion. Some parts of it are as follow:

       //data_set[t][8] = yaw = y, data_set[t][7] = pitch = x, data_set[t][6] = roll = z
        btQuaternion q(pcl::deg2rad(data_set[t][8]), pcl::deg2rad(data_set[t][7]), pcl::deg2rad(data_set[t][6]));

        new_pose.orientation.x = q.x();
        new_pose.orientation.y = q.y();
        new_pose.orientation.z = q.z();
        new_pose.orientation.w = q.w();

        cout << "new_pose.orientation.x = " << new_pose.orientation.x << endl;
        cout << "new_pose.orientation.y = " << new_pose.orientation.y << endl;
        cout << "new_pose.orientation.z = " << new_pose.orientation.z << endl;
        cout << "new_pose.orientation.w = " << new_pose.orientation.w << endl;

I have cross-checked with an on-line converter and found that the values returned somehow swapped and have sign(+ve and -ve) problem. The outputs are as below:

From my code:

new_pose.orientation.x = 0.0419219
new_pose.orientation.y = -0.0300053
new_pose.orientation.z = -0.184328
new_pose.orientation.w = 0.981512

From the on-line converter:

u0 = 0.9811347012447917 
u1 = -0.18632481923344862   
u2 = -0.04192188282716083   
u3 = -0.030005257724968114

I guess u0 corresponds to w, u1 corresponds to x and so on..

What is actually the problem of this conversion?