如何构建字典以实现堆叠DataFrame多列的便捷填充？

Great question! Let's break this down step by step to make filling your DataFrame as smooth as possible—no manual dictionary wrangling required once you set up the right mappings.

Step 1: Automate Match-to-Odds Mapping (No Manual Entry!) #

First, instead of manually creating that nested match dictionary (which is error-prone), we can write a simple function to convert your column-specific dictionary lists (like your LB_data) into a clean, reusable mapping. This also fixes any issues with inconsistent team order in your input dictionaries (e.g., if one dict has {'St Kilda': 3.3, 'Essendon': 1.32} but your DataFrame uses "Essendon v St Kilda" as the match key).

Here's the function:

def create_match_odds_map(odds_dicts):
    """Convert a list of team-odds dicts to a match-to-team-odds mapping."""
    match_map = {}
    for team_odds in odds_dicts:
        # Sort teams to ensure consistent match keys (e.g., "Essendon v St Kilda" always)
        sorted_teams = sorted(team_odds.keys())
        match_key = f"{sorted_teams[0]} v {sorted_teams[1]}"
        match_map[match_key] = team_odds
    return match_map

Use it for all your 4 columns:

# Your original LB data
LB_data = [
    {'Essendon': 1.32, 'St Kilda': 3.3}, {'Carlton': 5.0, 'Port Adelaide': 1.16},
    {'Geelong Cats': 1.57, 'Melbourne': 2.36}, {'Greater Western Sydney': 2.75, 'West Coast Eagles': 1.44},
    {'Brisbane': 1.95, 'North Melbourne': 1.85}, {'Hawthorn': 1.38, 'Western Bulldogs': 3.0},
    {'Fremantle': 1.32, 'Gold Coast': 3.3}
]

# Generate mappings for all 4 columns (replace Odds2_data/Odds3_data/Odds4_data with your actual data)
lb_map = create_match_odds_map(LB_data)
odds2_map = create_match_odds_map(Odds2_data)
odds3_map = create_match_odds_map(Odds3_data)
odds4_map = create_match_odds_map(Odds4_data)

Step 2: Fill Your DataFrame (Two Common Scenarios) #

Now, how you fill depends on your DataFrame's structure. Let's cover the two most likely cases:

Scenario 1: Your DataFrame has one row per team (stacked format) #

If your DataFrame looks like this (each row is a team in a match):

Use apply() to pull the right odds for each team and match:

import pandas as pd

# Replace df with your actual DataFrame
df['LB'] = df.apply(lambda row: lb_map.get(row['Match'], {}).get(row['Team']), axis=1)
df['Odds2'] = df.apply(lambda row: odds2_map.get(row['Match'], {}).get(row['Team']), axis=1)
df['Odds3'] = df.apply(lambda row: odds3_map.get(row['Match'], {}).get(row['Team']), axis=1)
df['Odds4'] = df.apply(lambda row: odds4_map.get(row['Match'], {}).get(row['Team']), axis=1)

Scenario 2: Your DataFrame has one row per match (with home/away teams) #

If your DataFrame has one row per match with separate columns for home/away teams:

Fill directly by referencing the match map and team columns:

# Fill LB column for home/away
df['LB_Home'] = df.apply(lambda row: lb_map[row['Match']][row['HomeTeam']], axis=1)
df['LB_Away'] = df.apply(lambda row: lb_map[row['Match']][row['AwayTeam']], axis=1)

# Repeat for other columns
df['Odds2_Home'] = df.apply(lambda row: odds2_map[row['Match']][row['HomeTeam']], axis=1)
df['Odds2_Away'] = df.apply(lambda row: odds2_map[row['Match']][row['AwayTeam']], axis=1)

Step 3: Optimize for Large Datasets (Optional) #

If you're working with a huge DataFrame, the apply() method might be slow. Instead, create a flattened mapping where the key is a tuple of (Match, Team) for faster lookups:

def create_flat_odds_map(match_map):
    flat_map = {}
    for match, team_odds in match_map.items():
        for team, odd in team_odds.items():
            flat_map[(match, team)] = odd
    return flat_map

# Create flattened maps for all columns
lb_flat = create_flat_odds_map(lb_map)
odds2_flat = create_flat_odds_map(odds2_map)

# Fill using tuple keys (much faster for large data)
df['LB'] = df.apply(lambda row: lb_flat.get((row['Match'], row['Team'])), axis=1)

Key Takeaways #

• Automate mappings: Never manually write match dictionaries—use the create_match_odds_map function to avoid typos and handle team order inconsistencies.
• Choose the right fill method: Use apply() for small datasets, or flattened mappings for large ones.
• Scale easily: This approach works for all 4 of your columns—just repeat the mapping and filling steps for each.

内容的提问来源于stack exchange，提问作者RJM