关于将H/T（抛硬币结果）转换为0/1二进制数组的技术咨询

Got it, let's get this sorted! You’ve loaded your sample data with Data = np.load('sample_1.npy') and need to turn those 'H'/'T' values into a 0/1 binary array. Here are a few efficient, straightforward methods using NumPy that should work depending on your data structure:

Method 1: Boolean Indexing (Most Efficient) #

This is the fastest approach for large datasets since it uses NumPy's vectorized operations under the hood:

# Initialize a zero array with the same shape as your data, integer type
binary_data = np.zeros_like(Data, dtype=int)
# Set all 'H' positions to 1 (swap 'H' with 'T' if you want T=1 instead)
binary_data[Data == 'H'] = 1

If you need to handle unexpected values (like something that’s not 'H' or 'T'), you can add a check to mark those as a placeholder (e.g., -1):

binary_data = np.zeros_like(Data, dtype=int)
binary_data[Data == 'H'] = 1
binary_data[Data == 'T'] = 0
# Mark non-H/T values as -1
binary_data[~np.isin(Data, ['H', 'T'])] = -1

Method 2: One-Liner with np.where #

For a concise, readable solution, use np.where to map values in one line:

# Map 'H' to 1, everything else (including 'T') to 0
binary_data = np.where(Data == 'H', 1, 0)
# Or reverse it: map 'T' to 1, 'H' to 0
binary_data = np.where(Data == 'T', 1, 0)

To handle invalid values here too, nest another np.where:

binary_data = np.where(Data == 'H', 1, np.where(Data == 'T', 0, -1))

Method 3: Vectorized Function (For Simple Use Cases) #

If you prefer a more explicit function-based approach, use np.vectorize (note: this is less efficient for large datasets than the above methods, but works fine for smaller samples):

def convert_flip(flip_result):
    if flip_result == 'H':
        return 1
    elif flip_result == 'T':
        return 0
    else:
        return -1  # Handle invalid entries

# Create a vectorized version of the function
vectorized_convert = np.vectorize(convert_flip)
binary_data = vectorized_convert(Data)

Just make sure your Data array is a string dtype (you can check with Data.dtype). If it’s stored as object dtype, the above methods still work—NumPy will handle the string comparisons correctly.

内容的提问来源于stack exchange，提问作者PythonBeginner