英文:
How to skip a row of zeros with np.loadtxt?
问题
我有一个包含许多行数值的.txt文件。每一行看起来如下
3.896784 0.465921 1.183185 5.468042 ...，其中数值根据行不同而异。
此外，每一行都包含900个数值。
有些行不包含真实数据，只包含900个零，如下所示
0 0 0 0 0 0 0 0 ...。
我知道如何跳过只包含一个值的行，即一个零，例如通过以下代码：
import numpy as np
data = np.loadtxt("pathtodata")
data = data[~(data==0)]
但是这段代码无法处理每行包含2个或更多零值的情况。是否有一种方法可以不加载只包含900个或任意数量的零（或特定整数）的行？
英文:
I have a .txt file that contains many rows of numeric values. Each row looks as follows
3.896784 0.465921 1.183185 5.468042 ..., where the values differ depending on the row.
Furthermore, every row contains 900 values.
Some rows do not contain real data, but only 900 zeros as follows
0 0 0 0 0 0 0 0 ....
I know how to skip rows that only contain one value, i.e., one zero, such as via the following code:
import numpy as np
data = np.loadtxt("pathtodata")
data = data[~(data==0)]
But this code does not work for 2 zero values or more per row. Is there a way to not load rows that only contain 900 or any arbitrary number of zeros (or a specific integer)?
答案1
得分: 2
import numpy as np
data = np.loadtxt("data.txt")
data = data[~np.all(data == 0, axis=1)]
for row in data:
print(row)
这应该可以解决问题，np.all(data == 0, axis=1) 本质上就是你所要求的，它检查每行中的每个元素是否满足条件，在这里你排除了每个元素都等于0的行。
英文:
import numpy as np
data = np.loadtxt("data.txt")
data = data[~np.all(data == 0, axis=1)]
for row in data:
print(row)
Should do the trick, np.all(data == 0, axis=1) is essentially what you are asking for, it checks if every element in a given row verify a condition, here you exclude every row where every element are equal to 0
答案2
得分: 1
你可以使用 np.any 来实现这个功能。它会检查沿着特定轴是否有非零值。所以当一行全是零时，它会返回 False。然后你可以使用这个输出来对数据集进行子集化：
import numpy as np
test = np.arange(20).reshape((4, -1))
test[0, :] = 0
print(test)
mask = np.any(test, axis=1)
print(test[mask, :])
英文:
You can use np.any to do that. It will check if there are any values other than 0 along a certain axis. So it will return False when a row is all zeros. Then you can use the ouptut of this to subset your dataset:
import numpy as np
test = np.arange(20).reshape((4,-1))
test[0,:] = 0
print(test)
mask = np.any(test, axis=1)
print(test[mask,:])
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库，让每个人都能够通过互相帮助和分享经验来进步。
评论