Python Notes14:HDF5相关操作

数据读取

import h5py
import numpy as np

out_h5_path = r'D:\DataDemo\test_hdf5.h5'

f = h5py.File(out_h5_path, 'r')
f.keys()  # 可以查看所有的主键
print([key for key in f.keys()])

# 根据key中的字符来提取h5文件中的值
image_stack = f['train'][:]
print(image_stack)
print(image_stack.dtype)
print(image_stack.shape)
# f.close()
print(image_stack[1])

创建

• group

• dataset

import h5py

# 中文路径读不出来
out_h5_path = r'D:\DataDemo\demo1.h5'

with h5py.File(out_h5_path, 'w') as f:
    # f.create_group()
    # f.create_group("group1")
    f.create_dataset("group/dataset",(10,), dtype='i')

h5py文档

Quick Start Guide - h5py 3.5.0 documentation

With Anaconda or Miniconda: If there are wheels for your platform (mac, linux, windows on x86) and you do not need MPI you can install via pip: With Enthought Canopy, use the GUI package manager or: To install from source see Installation . "HDF" stands for "Hierarchical Data Format".

https://docs.h5py.org/en/latest/quick.html