將新數據添加到 HDF5 文件會導致空數組

Question

在玩 Python 的 HDF5 包時，我發現了一個奇怪的行為。 我想在表中插入更多數據。 但不知何故我無法讓它正常工作。 正如您從源代碼中看到的那樣，我使用fromRow = hf["X"].shape[0]獲取鍵 'X' 中的最后一行數據，然后寫入tempArray2 。 結果是一個空表。

import h5py

tempArray1 = [[0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443], [0.9293237924575806, -0.32789671421051025, 0.18110771477222443]]
tempArray2 = [[3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14], [3.1387749004352372e-06, 8.120089097236803e+27, -1.645612730013634e-14]]

with h5py.File('data.hdf5', 'w') as hf:
    # Add data to new file
    dset = hf.create_dataset("X", data=tempArray1, compression="gzip", chunks=True, maxshape=(None,3), dtype='f4') # Size is as the size of tempArray1
    print(hf["X"].shape[0])

    # Append data existing file
    hf["X"].resize((hf["X"].shape[0] + 10, 3)) # Size is as the size of X+ 10
    print(hf["X"].shape[0])
    fromRow = hf["X"].shape[0]
    hf["X"][fromRow:] = tempArray2

這是它的外觀：

Key: X
Data:
 [[ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.9293238  -0.3278967   0.18110771]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]
 [ 0.          0.          0.        ]]
Length of data: 20

奇怪的是，當我用數字 10 替換值fromRow ，例如fromRow = 10 ，它代表現有表的結尾，它起作用了。

輸出：

Key: X
Data:
 [[ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 9.2932379e-01 -3.2789671e-01  1.8110771e-01]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]
 [ 3.1387749e-06  8.1200891e+27 -1.6456127e-14]]
Length of data: 20

知道我做錯了什么嗎？

Answer 1

調整 X 數據集大小后，您將獲得fromRow 。 您需要在調整大小之前的值。 請參閱下面的代碼。

with h5py.File('data.hdf5', 'w') as hf:
    # Add data to new file
    dset = hf.create_dataset("X", data=tempArray1, compression="gzip", chunks=True, maxshape=(None,3), dtype='f4') # Size is as the size of tempArray1
    print(hf["X"].shape[0])
# new location to get fromRow:
    fromRow = hf["X"].shape[0]

    # Append data existing file
    hf["X"].resize((hf["X"].shape[0] + 10, 3)) # Size is as the size of X+ 10
    print(hf["X"].shape[0])        
    hf["X"][fromRow:] = tempArray2