用h5py一次写一个多维结构numpy数组一个字段引发一个numpy广播错误

Question

我想弄清楚如何用h5py每次写一个hdf5文件一个字段的多维结构numpy数组，但是我收到有关播放不同形状阵列的错误。我需要分离创建数据集的步骤，并将数据写入数据集，如示例中所示。

这是我最小的工作示例。

writehdf5.py

#!/bin/env python 
import h5py 
from numpy.random import randn 

print 'Creating Test Data' 
mass = randn(10) 
altitude = randn(10) 
position = randn(10, 3) 
velocity = randn(10, 3) 

print 'Write 1 dimensional arrays' 
hdf5 = h5py.File('test1.hdf', 'w') 
dataset = hdf5.create_dataset('test dataset', (10,), 
           dtype=[('mass', '<f8'), 
            ('altitude', '<f8')]) 
dataset['mass'] = mass 
dataset['altitude'] = altitude 
hdf5.close() 

print 'Write 2 dimensional arrays' 
hdf5 = h5py.File('test2.hdf', 'w') 
dataset = hdf5.create_dataset('test dataset', (10,), 
           dtype=[('position', '<f8', 3), 
            ('velocity', '<f8', 3)]) 
print dataset['position'].shape 
print position.shape 
dataset['position'] = position # <-- Error Occurs Here 
dataset['velocity'] = velocity 
hdf5.close() 

，当运行给我下面的输出。

>> python writehdf5.py 
Creating Test Data 
Write 1 dimensional arrays 
Write 2 dimensional arrays 
(10, 3) 
(10, 3) 
Traceback (most recent call last): 
    File "mwe.py", line 27, in <module> 
    dataset['position'] = position # <-- Error Occurs Here 
    File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper (/builddir/build/BUILD/h5py-2.5.0/h5py/_objects.c:2450) 
    File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper (/builddir/build/BUILD/h5py-2.5.0/h5py/_objects.c:2407) 
    File "/usr/lib64/python2.7/site-packages/h5py/_hl/dataset.py", line 514, in __setitem__ 
    val = numpy.asarray(val, dtype=dtype, order='C') 
    File "/usr/lib64/python2.7/site-packages/numpy/core/numeric.py", line 462, in asarray 
    return array(a, dtype, copy=False, order=order) 
ValueError: could not broadcast input array from shape (10,3) into shape (10,3,3) 

>> h5dump test1.hdf 
HDF5 "test1.hdf" { 
GROUP "/" { 
    DATASET "test dataset" { 
     DATATYPE H5T_COMPOUND { 
     H5T_IEEE_F64LE "mass"; 
     H5T_IEEE_F64LE "altitude"; 
     } 
     DATASPACE SIMPLE { (10)/(10) } 
     DATA { 
     (0): { 
      0.584402, 
      1.50107 
     }, 
     (1): { 
      -0.284148, 
      -0.521783 
     }, 
     (2): { 
      -0.461751, 
      0.53352 
     }, 
     (3): { 
      2.06525, 
      -0.0364377 
     }, 
     (4): { 
      -0.835377, 
      1.35912 
     }, 
     (5): { 
      -1.31011, 
      1.21051 
     }, 
     (6): { 
      0.103971, 
      -0.669617 
     }, 
     (7): { 
      0.244425, 
      -0.654791 
     }, 
     (8): { 
      0.468478, 
      2.60204 
     }, 
     (9): { 
      0.837614, 
      1.21362 
     } 
     } 
    } 
} 
} 

>> h5dump test2.hdf 
HDF5 "test2.hdf" { 
GROUP "/" { 
    DATASET "test dataset" { 
     DATATYPE H5T_COMPOUND { 
     H5T_ARRAY { [3] H5T_IEEE_F64LE } "position"; 
     H5T_ARRAY { [3] H5T_IEEE_F64LE } "velocity"; 
     } 
     DATASPACE SIMPLE { (10)/(10) } 
     DATA { 
     (0): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (1): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (2): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (3): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (4): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (5): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (6): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (7): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (8): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     }, 
     (9): { 
      [ 0, 0, 0 ], 
      [ 0, 0, 0 ] 
     } 
     } 
    } 
} 
} 

我可以看到我的数据集被正确初始化，但我不明白，当我试图填补与数据的字段，我得到的错误。数据集和要写入的数据的形状明显相同。

对我所缺少的任何帮助（很可能是简单的东西）非常感谢！

Answer 1

您的脚本在我的机器（PY3），h5py版本2.6.0上正常运行。看起来你有2.5。

它必须是v2.5.0中的一个bug，它已在v2.6.0中修复。