Python netCDF4使用Dataset的memory参数，提高对较大.nc文件读取效率

Python netCDF4通过一次性读取到内存，提高对较大.nc文件读取效率

对大量的.nc处理瓶颈卡在读与写上，全是机械硬盘就只能考虑改代码了。发现netCDF4中Dataset的构造函数中有memory参数，默认为None，如果没有设置memory，则访问数据的时候从硬盘中读取，若有，则从内存中读取。

import netCDF4 as nc # netcdf4 1.5.7
from time import perf_counter

# 不一次性读取到内存上
time1 = 0
for i in range(10):
    start_time = perf_counter()
    dataset = nc.Dataset(r'H:\test.nc')
    for key in dataset.variables.keys():
        data = dataset[key][:]
    dataset.close()
    end_time = perf_counter()
    time1 += end_time - start_time

# 一次性读取到内存上
time2 = 0
for i in range(10):
    start_time = perf_counter()
    file = open(r'H:\test.nc','rb')
    memory = file.read()
    dataset = nc.Dataset(r'H:\test.nc',mode='r',memory=memory)
    for key in dataset.variables.keys():
        data = dataset[key][:]
    dataset.close()
    file.close()
    # 如果不需要了可以提前释放内存
    memory = None
    end_time = perf_counter()
    time2 += end_time - start_time