如何在python中更新现有的netCDF数组

我有一个如下所示的代码，我想根据另一个NetCDF文件更新现有的NetCDF文件。如果在现有NetCDF文件的另一个NetCDF文件

value=0

，则该文件为0。所以我只考虑其他NETCDF不等于0

的地方。

import netCDF4 as net

days=["01","02","03","04","05","06","07","08","09","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31"]
months=["02","03","04","05","06","07"]

for month in months:
    for day in days:
        try:
            infile=net.Dataset("2015%s%s-160000_0.nc" % (month,day),"w")
            value=infile.variables['HailProb'][:]
            value1=infile1.variables['hailsie'][:]
            infile[value==0 or value1==0]==0
            infile.close()
        except RuntimeError:
            continue

---

因为如果我在写模式下打开

infle

，我担心所有现有数据都会被重写。有人能看看我是如何解决这个问题的吗？代码有什么问题？

用

a

以追加模式打开文件：

infile=net.Dataset("2015%s%s-160000_0.nc" % (month,day), "a")

现在，您可以读取和写入现有文件，如文档所述：

a

和

r+

平均附加（与 串行文件）；打开现有文件进行读写操作

---

我相信你的代码有几个问题

为了更改现有文件，我认为您应该在

r+

（或者，如Mike Müller所述，在

a

）模式下阅读它

对于

infle[value==0或value1==0]==0

，

infle

不是您的数据数组，而是

netCDF4

对象。要更改变量，您需要类似于

infle.variables[“HailProb”]…

使用

infle[..]==0

，

=

测试相等性，它不会在

infle

中设置新值（如果

infle

实际上是一个数组，则使用

=

）

使用

value==0或value1==0的行将不起作用；需要类似于
（value==0）|（value1==0）
的内容来切片数组

我认为这样做应该奏效：

import netCDF4 as net

days=["01","02","03","04","05","06","07","08","09","10","11","12","13","14","15","16","17","18","19","20","21","22","23","24","25","26","27","28","29","30","31"]
months=["02","03","04","05","06","07"]

for month in months:
    for day in days:
        try:
            infile = net.Dataset("2015%s%s-160000_0.nc" % (month,day), "r+")
            value  = infile.variables['HailProb'][:]
            value1 = infile1.variables['hailsie'][:]
            infile.variables['HailProb'][(value==0) | (value1==0)] = 0
            infile.close()
        except RuntimeError:
            continue

一个更完整的工作示例，它还创建了NetCDF文件以供使用：

import netCDF4 as nc4
import numpy as np

# Create NetCDF file 1
nc1  = nc4.Dataset('nc1.nc', 'w')
nc1.createDimension('z', 10)
nc_var = nc1.createVariable('HailProb', 'f8', 'z')

# Add some dummy data, and set a single value to zero:
nc_var[:] = np.random.randint(2, 10, size=10)
nc_var[5] = 0
nc1.close()

# Create NetCDF file 2
nc2  = nc4.Dataset('nc2.nc', 'w')
nc2.createDimension('z', 10)
nc_var = nc2.createVariable('hailsie', 'f8', 'z')

# Add some dummy data, and set a single value to zero:
nc_var[:] = np.random.randint(2, 10, size=10)
nc_var[3] = 0
nc2.close()

# Read files back
# File 1 in in "r+" mode, since we want to edit it
nc1 = nc4.Dataset('nc1.nc', 'r+')
value = nc1.variables["HailProb"][:]

# File 2 only required "r" (read) mode
nc2 = nc4.Dataset('nc2.nc', 'r')
value1 = nc2.variables["hailsie"][:]

# Update NetCDF file based on some conditions
nc1.variables["HailProb"][(value==0) | (value1==0)] = 0

nc1.close()
nc2.close()

运行代码的第一部分（创建

nc1.nc

和

nc2.nc

）后，

ncdump-v HailProp nc1.nc

和

ncdump-v hailcie nc2.nc

给出了以下信息：

HailProb = 2, 9, 6, 6, 9, 0, 5, 4, 8, 6 ;
hailsie  = 4, 6, 2, 0, 4, 3, 6, 4, 5, 7 ;

运行代码的第二部分后，我得到：

HailProb = 5, 6, 8, 0, 7, 0, 6, 4, 2, 4 ;

---

追加模式不是只从数组末尾追加而不是更新吗？现在的问题是infle[value==0或value1==0]==0是错误的，它给了我一个错误。你想用

infle[value==0或value1==0]==0实现什么。不能将
[]
与文件对象一起使用。[（值==0）|（值1==0）]isnit条的含义是什么？但实际上我想要一个orNo，
&
意味着，
是或。例如，参见文档
numpy
中也有内置的运算符，因此类似于
nc1.variables[“HailProb”][np.logical_或（value==0，value1==0）]=0的东西也应该像我说的那样工作。该文件根本没有被更新。它仍然是同一个文件，我可以看到填充的日期没有改变。你能更新你的帖子并添加你现在正在尝试的代码吗？否则，我们只能猜测到底出了什么问题