如何重新定义int64数组变量：筛选有效值并排序

Got it, let's work through this to get the exact result you're looking for. Here are two approaches—one that builds on your existing code, and a more streamlined version that cuts out unnecessary steps.

Approach 1: Build on Your Existing Code #

First, let's fix a small variable naming conflict in your original code, then add the filtering and sorting step:

import numpy as np

# Original raw data
data = ['10 20 10 36 30 33 400 400 -1 -1', 
        '100 50 50 30 60 27 70 24 -2 -2 700 700', 
        '300 1000 80 21 90 18 100 15 110 12 120 9 900 900 -3 -3', 
        '30 90 130 6 140 3 -4 -4 1000 1000']

# Split each string into individual elements
data_split = [e.split() for e in data]

# Extract concentration values (every 2nd element starting at index 3) and convert to numpy arrays
concentration = [np.array(row[3::2], dtype=np.int) for row in data_split]

# Step 1: Mark out-of-range values as 0 (your existing logic)
for row_idx in range(len(concentration)):
    for elem_idx in range(len(concentration[row_idx])):
        val = concentration[row_idx][elem_idx]
        if not (0 <= val <= 50):
            concentration[row_idx][elem_idx] = 0
            print(f"Error: Index {elem_idx} in row {row_idx} is out of range")

# Step 2: Filter out 0s and sort the valid values
concentration_clean_sorted = [np.sort(arr[arr != 0]) for arr in concentration]

# Print the final result
for arr in concentration_clean_sorted:
    print(f"Array of int64 {arr}")

How This Works: #

• We use numpy boolean indexing (arr != 0) to quickly filter out all invalid 0 values from each array.
• np.sort() sorts the remaining valid values in ascending order in one line—far cleaner than writing nested loops for sorting.

Approach 2: Streamlined (No Intermediate 0 Step) #

If you want to skip marking values as 0 entirely and go straight to filtering valid values, this version is more efficient:

import numpy as np

data = ['10 20 10 36 30 33 400 400 -1 -1', 
        '100 50 50 30 60 27 70 24 -2 -2 700 700', 
        '300 1000 80 21 90 18 100 15 110 12 120 9 900 900 -3 -3', 
        '30 90 130 6 140 3 -4 -4 1000 1000']

data_split = [e.split() for e in data]
concentration_clean_sorted = []

for row in data_split:
    # Extract and convert concentration values to numpy array
    arr = np.array(row[3::2], dtype=np.int)
    
    # Filter values that are within 0-50 range
    valid_vals = arr[(arr >= 0) & (arr <= 50)]
    
    # Sort the valid values
    sorted_vals = np.sort(valid_vals)
    concentration_clean_sorted.append(sorted_vals)
    
    # Print error messages for out-of-range indices (optional)
    invalid_indices = np.where((arr < 0) | (arr > 50))[0]
    for idx in invalid_indices:
        print(f"Error: Index {idx} in current row is out of range")

# Output the final sorted arrays
for arr in concentration_clean_sorted:
    print(f"Array of int64 {arr}")

Why This Is Better: #

• It combines validation, filtering, and sorting into a single pass, avoiding the need to modify the original arrays with 0s first.
• Uses numpy's vectorized operations which are faster than nested Python loops, especially for larger datasets.

Final Output #

Both approaches will give you exactly the result you want:

Array of int64 [33 36]
Array of int64 [24 27 30]
Array of int64 [12 15 18 21]
Array of int64 [3 6]

内容的提问来源于stack exchange，提问作者Kimdey