如何更新字典：更新已有值并追加新值（含DataFrame场景）

Great question! Let’s break this down into two clear parts: first the general dictionary update logic, then the specific pandas DataFrame scenario you’re working through.

General Dictionary Update Logic #

At its core, updating a dictionary to sum existing values and add new entries just requires checking each key from your new data: if the key exists, add the new value to the existing one; if not, create a new key with the new value. Here are two straightforward ways to do this:

Method 1: Explicit Loop with dict.get() #

This is easy to follow, even if you’re new to Python:

my_dict = {'key1': 21, 'key2': 15}
new_data = {'key1': 18, 'key3': 7}

for key, value in new_data.items():
    # Use get() to safely retrieve existing value (or 0 if key doesn't exist)
    my_dict[key] = my_dict.get(key, 0) + value

print(my_dict)  # Output: {'key1': 39, 'key2': 15, 'key3': 7}

Method 2: Using collections.Counter #

If you’re working with numeric counts (like your use case), Counter from the collections module is purpose-built for this kind of aggregation:

from collections import Counter

my_counter = Counter({'key1': 21, 'key2': 15})
new_data = Counter({'key1': 18, 'key3': 7})

# The update() method automatically sums existing keys and adds new ones
my_counter.update(new_data)

print(dict(my_counter))  # Output: {'key1': 39, 'key2': 15, 'key3': 7}

Your Specific Pandas DataFrame Scenario #

Now let’s apply this to your task: aggregating value_counts() results from multiple DataFrames into one dictionary, with summed counts for matching keys and new keys for unique entries.

Solution 1: Using Counter (Simplest Approach) #

Since value_counts() returns a pandas Series that converts cleanly to a Counter, this is the most concise method:

import pandas as pd
from collections import Counter

# Initialize an empty Counter to hold all aggregated counts
total_counts = Counter()

# Replace this with your actual list of N DataFrames
dataframes = [
    pd.DataFrame({'col': ['key1', 'key1', 'key2']}),
    pd.DataFrame({'col': ['key1', 'key3', 'key3', 'key3']}),
    pd.DataFrame({'col': ['key2', 'key2', 'key4']})
]

# Loop through each DataFrame and update the total counts
for df in dataframes:
    df_counts = Counter(df['col'].value_counts().to_dict())
    total_counts.update(df_counts)

# Convert back to a regular dictionary if needed
final_dict = dict(total_counts)
print(final_dict)
# Output: {'key1': 3, 'key2': 3, 'key3': 3, 'key4': 1}

Solution 2: Using a Regular Dictionary (No Extra Imports) #

If you prefer not to use collections, you can use the dict.get() method from earlier:

import pandas as pd

# Initialize empty dictionary for aggregated counts
total_counts = {}

dataframes = [
    pd.DataFrame({'col': ['key1', 'key1', 'key2']}),
    pd.DataFrame({'col': ['key1', 'key3', 'key3', 'key3']}),
    pd.DataFrame({'col': ['key2', 'key2', 'key4']})
]

for df in dataframes:
    # Get value counts as a dictionary
    df_counts = df['col'].value_counts().to_dict()
    # Update the total dictionary
    for key, count in df_counts.items():
        total_counts[key] = total_counts.get(key, 0) + count

print(total_counts)
# Same output: {'key1': 3, 'key2': 3, 'key3': 3, 'key4': 1}

Both methods work perfectly for your scenario—pick whichever you find more readable!

内容的提问来源于stack exchange，提问作者Vash