Pandas calculates the distance between two longitudes and latitudes. The amount of data is too large, hundreds of millions of lines. Is there an efficient algorithm?

The phenomenon and background of the problem

pandas Calculate the distance between two longitudes and latitudes , Using methods that can be found on the Internet , I haven't seen the progress for half a day

Problem related code , Do not paste screenshots

lon1, lat1, lon2, lat2 = map(np.radians, tt.tolist()) # haversine The formula dlon = lon2 - lon1dlat = lat2 - lat1a = np.sin(dlat / 2) ** 2 + np.cos(lat1) * np.cos(lat2) * np.sin(dlon / 2) ** 2c = 2 * a * np.sin(np.sqrt(a))r = 6371c * r * 1000

Operation results and error reporting contents

Running time is too long

My solution ideas and tried methods

no

What I want to achieve

Efficient algorithm

Now that we have used the available methods, we can not reduce the operation time , It is no longer a problem of algorithm optimization , So your problem is that the computing performance of the machine is not enough .
If the computer's graphics card is good, it is recommended to use minpy Speed up the operation , If the computer graphics card is not good or there is no independent graphics card , It is recommended to apply for calculation .
notes ：python There may not be any in computing speed C/C++ fast , You can try another language .