Out[64]: CPU times: user 2 ms, sys: 216 ms, total: 218 ms
Wall time: 216 ms
array([[ 0.10989742, -0.48626177, -0.60849881, ..., -0.99051776,
0.88124291, -1.34261656],
[-0.42301145, 0.29831708, 1.29729826, ..., -0.73426192,
-0.13484905, 0.91787421],
[ 0.12322789, -0.28728811, 0.85956891, ..., 1.47888978,
-1.12452641, -0.528133 ],
...,
[ 0.06507559, -0.37130379, 1.35427048, ..., -1.4457718 ,
0.49509821, 0.0738847 ],
[ 1.76525714, -0.07876135, -2.94133788, ..., -0.62581084,
0.0933164 , 1.55788205],
[-1.18439949, -0.73210571, -0.45845113, ..., 0.0528656 ,
-0.39526633, -0.5964333 ]])
In [ 65 ]: data = 0.0
!rm -f $path*
In any case, you can expect that this form of data storage and retrieval is much, much
faster as compared to SQL databases or using the standard pickle library for serialization.
Of course, you do not have the functionality of a SQL database available with this
approach, but PyTables will help in this regard, as subsequent sections show.
