如何在Python 3或Pandas中优化列表/DataFrame列的重复整数（保留重复ID）并生成期望列表？

Got it, let's solve this problem where you need to map duplicate IDs in a list (or DataFrame column) to sequential integers—all while keeping the duplicate occurrences intact. Below are two practical approaches, depending on whether you prefer pure Python or using Pandas:

If you're working with a regular list and don't want to import external libraries, a dictionary-based mapping works perfectly. We'll track each unique ID and assign it a sequential number as we iterate through the list:

original_array = [1, 1, 5, 8, 8, 20213, 22170, 22170]
id_mapping = {}
current_sequence = 1
result_array = []

for id_val in original_array:
    # Assign a new sequence number if the ID hasn't been seen before
    if id_val not in id_mapping:
        id_mapping[id_val] = current_sequence
        current_sequence += 1
    # Append the mapped number to the result
    result_array.append(id_mapping[id_val])

print(result_array)  # Output: [1, 1, 2, 3, 3, 4, 5, 5]

How it works: #

• We use id_mapping to store each unique original ID and its corresponding sequential number.
• Every time we encounter a new ID, we assign it the next available sequence number and increment the counter.
• Duplicate IDs reuse their already assigned number, so duplicates are preserved in the result.

If you're working with a Pandas DataFrame (which is common for tabular data with ID columns), the factorize() method is a built-in, efficient solution. It's designed exactly for this kind of categorical-to-sequential mapping:

import pandas as pd

# Sample DataFrame with the original ID column
df = pd.DataFrame({
    'original_id': [1, 1, 5, 8, 8, 20213, 22170, 22170]
})

# Factorize the column (default starts at 0, so we add 1 to start from 1)
df['new_id'] = pd.factorize(df['original_id'])[0] + 1

# Convert the result back to a list if needed
print(df['new_id'].tolist())  # Output: [1, 1, 2, 3, 3, 4, 5, 5]

How it works: #

• pd.factorize() returns two values: an array of integer codes for each element (matching the order of first occurrence), and an array of unique original values.
• We take the first returned array and add 1 to shift the sequence from starting at 0 to starting at 1 (remove the +1 if you want 0-based numbering).
• This method is optimized for large datasets, making it faster than manual loops for big DataFrames.

内容的提问来源于stack exchange，提问作者Wruktarr