As far as I know there is no fully automatic way in blender to do what you want to do.
The tracking information is unique for each tracked video (or image sequence) and cannot be shared.
As a workaround you can use the Orientation settings, it should not be difficult to match different shots if you prepare your scene carefully.
To understand how this works:
Before creating the solved scene, Blender allows you to determine the scale and orientation. When you hit the Setup Tracking Scene button blender then re-creates the relationship between tracked points and camera in the 3D viewport as a set of points parented to a camera according to the information provided by the scale and orientation settings.
The key is to set these parameters meticulously for each of the tracked shots you want to reconstruct.
To set scale
With a measuring tape, measure the distance between the points that you will be using as trackers in the real world.
Once you have tracked your shot (and before pressing the Setup Tracking Scene button) shift+select the trackers that correspond to those you've measured in the real world and enter the distance information on the Distance parameter and press Apply Scale.
Set a common point as origin.
Choose a point that is present in the different scenes to be tracked to work as origin. Choose the tracker and Press the Set Origin button.
On the new 3D scene, that tracking point will be placed at the center of the world at coordinates X:0 Y:0 Z:0. Doing that in all of the tracked video clips will allow you to match a common point in the scene.
Set X axis and Y axis
Select a tracker that is aligned with the X axis (with respect to the origin) then press Set X axis. Do the same for one a different tracking point that is aligned with the Y axis.
Once you've set the orientation and scale correctly, hit the Setup Tracking Scene button. If your settings for Sensor, Lens and Lens distortion are correct, and you've manage to track the scene with a low solve error (less than 0.3 pixels). You should be able to line up different shots in a common space and scale.
You can also do some manual adjusting if the automatic tools fail. Remember that the camera solution means that all of the points and camera are parented, so if you move one element the rest will follow. See: Correct tracked camera orientation
Additional info: How can I get better results when doing camera motion tracking?