Read and writing performance test between numpy.ndarray (.npy) vs obspy.Stream (.mseed).

ndarray_as_tmp_file

Python 3.6.4 |Anaconda custom (64-bit)| (default, Mar 12 2018, 20:05:31) 
Type 'copyright', 'credits' or 'license' for more information
IPython 6.2.1 -- An enhanced Interactive Python. Type '?' for help.

In [1]: import numpy as np

In [2]: import obspy

In [3]: data = np.random.randn(6000, 3, 2000)

In [4]: traces = []
   ...: for i, frame in enumerate(data):
   ...:     for j, trace in enumerate(frame):
   ...:         traces.append(obspy.Trace(data=trace))
   ...: st = obspy.Stream(traces)
   ...: 

In [5]: %timeit np.save("data.npy", data)
398 ms ± 3.96 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

In [6]: %timeit data = np.load("data.npy")
188 ms ± 868 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

In [7]: %timeit st.write("data.mseed")
2.24 s ± 48.1 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

In [8]: %timeit st = obspy.read("data.mseed")
3.09 s ± 78.5 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)