如何在循环中扩展NumPy数组？并将姓名添加到出勤数组最左侧

Hey there! Let's work through your two NumPy questions using the dataset you provided. I'll break each one down with practical examples that fit your specific use case.

First, a quick heads-up: NumPy arrays are designed to be fixed-size for performance. Extending them in a loop (like you might with a Python list) can be inefficient because each extension requires copying the entire array to a new memory block. But if you need to do it, here's how—and a better alternative for most cases.

Option 1: Dynamic extension in a loop (use sparingly) #

You can use np.vstack() (for rows) or np.hstack() (for columns) to append new elements in each iteration. Let's simulate adding 3 new attendance rows to your existing array:

import numpy as np

# Your existing attendance array
attendance = np.array([[1, 0, 0, 1, 1, 1, 0],
                       [1, 0, 1, 1, 1, 1, 1],
                       [1, 1, 1, 1, 1, 1, 0],
                       [1, 0, 0, 1, 1, 1, 1],
                       [1, 1, 1, 0, 1, 1, 1],
                       [1, 0, 0, 0, 1, 1, 1],
                       [1, 1, 1, 1, 1, 1, 1],
                       [1, 1, 0, 0, 0, 0, 0],
                       [1, 1, 1, 1, 1, 1, 1],
                       [0, 0, 0, 1, 1, 1, 1]])

# Initialize empty array with matching column count
extended_arr = np.array([], dtype=int).reshape(0, 7)

# Add existing rows first (or start empty if building from scratch)
extended_arr = np.vstack([extended_arr, attendance])

# Loop to add new rows (simulating dynamic data)
for _ in range(3):
    # Generate a random attendance row for example
    new_row = np.random.randint(0, 2, size=7)
    extended_arr = np.vstack([extended_arr, new_row])

print(extended_arr.shape)  # Output: (13, 7) — 10 original + 3 new rows

Option 2: Pre-allocate the array (far more efficient) #

If you know the final size of your array upfront, pre-allocate memory first, then fill it in the loop. This avoids repeated array copies:

# We know we want 13 total rows (10 original + 3 new)
total_rows = 13
pre_allocated = np.empty((total_rows, 7), dtype=int)

# Fill existing data first
pre_allocated[:10] = attendance

# Loop to fill new rows
for i in range(10, total_rows):
    pre_allocated[i] = np.random.randint(0, 2, size=7)

print(pre_allocated.shape)  # Output: (13, 7)

This is the preferred method for large datasets or long loops—it's way faster and uses memory more efficiently.

Since name_list contains strings and attendance contains integers, we can't combine them directly into a standard NumPy array (which requires homogeneous data types). Instead, we have two good options:

Option 1: Use an object-type array (mixed data types) #

Convert attendance to an object-type array (which can hold any Python object, including strings and integers), then stack the names as a new left column:

name_list = np.array(["Ali","Ahmad","Beng","Chris","Sita","Marion","Stephen","Cobby","Akmal","Nita"])

# Reshape names into a 2D array (1 column, 10 rows)
name_col = name_list.reshape(-1, 1)
# Convert attendance to object type to support mixing with strings
attendance_obj = attendance.astype(object)
# Horizontally stack the name column and attendance data
final_array = np.hstack([name_col, attendance_obj])

# Print the result
for row in final_array:
    print(row)

Sample output:

['Ali' 1 0 0 1 1 1 0]
['Ahmad' 1 0 1 1 1 1 1]
['Beng' 1 1 1 1 1 1 0]
...

Option 2: Use a structured array (tabular data with named columns) #

If you want a more structured, table-like format (with named columns for each day and the name), use a structured array. This lets you access columns by name (e.g., result['name'] or result['Sunday']):

day_list = np.array(["Sunday","Monday","Tuesday","Wednesday","Thursday","Friday","Saturday"])

# Define the data type: name (string) + one integer column per day
dtype_spec = [('name', 'U10')] + [(day, int) for day in day_list]
# Create an empty structured array with the right length
structured_result = np.empty(len(name_list), dtype=dtype_spec)

# Fill in the names
structured_result['name'] = name_list
# Fill in attendance data for each day
for idx, day in enumerate(day_list):
    structured_result[day] = attendance[:, idx]

# Access data by column name
print("Names:", structured_result['name'])
print("Sunday attendance:", structured_result['Sunday'])

This is great if you plan to work with the data like a table, as it keeps each column's type clear and accessible.

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