Sort by Column - Pandas Step by Step

Pick the minimum-score row from the remaining pool repeatedly until the pool is empty, collecting names in order. With pandas, df.sort_values('score') returns a new DataFrame with all rows in ascending order by that column.

By hand

Extract scores and names as parallel lists. Keep a remaining index pool. Each iteration picks the index of the minimum remaining score, appends the matching name to result, and removes that index from the pool.

                            naive.py
                        
Replay: real traced execution (multi-file project)

                            # trace: ignore records
records = [
    {'name': 'al', 'score': 80},
    {'name': 'bo', 'score': 55},
    {'name': 'cy', 'score': 90},
    {'name': 'di', 'score': 70},
]
scores = [r['score'] for r in records]
names = [r['name'] for r in records]
remaining = list(range(len(records)))
result = []
for _ in range(len(records)):
    best = min(remaining, key=lambda i: scores[i])
    result.append(names[best])
    remaining.remove(best)
print('RESULT:', result)

                        
1{'name': 'al', 'score': 80},⋯2records = [3    {'name': 'al', 'score': 80},4    {'name': 'bo', 'score': 55},⋯
2{'name': 'bo', 'score': 55},⋯3{'name': 'al', 'score': 80},4{'name': 'bo', 'score': 55},5{'name': 'cy', 'score': 90},⋯
3{'name': 'cy', 'score': 90},⋯4{'name': 'bo', 'score': 55},5{'name': 'cy', 'score': 90},6{'name': 'di', 'score': 70},⋯
4{'name': 'di', 'score': 70},⋯5    {'name': 'cy', 'score': 90},6    {'name': 'di', 'score': 70},7]⋯
5records = [1# trace: ignore records2records = [3    {'name': 'al', 'score': 80},⋯
6scores ← [80, 55, 90, 70]⋯7]8scores = [r['score'] for r in records]9names = [r['name'] for r in records]⋯values this step[80, 55, 90, 70]scores
7names ← ['al', 'bo', 'cy', 'di']⋯8scores = [r['score'] for r in records]9names = [r['name'] for r in records]10remaining = list(range(len(records)))⋯values this step['al', 'bo', 'cy', 'di']names
8remaining ← [0, 1, 2, 3]⋯9names = [r['name'] for r in records]10remaining = list(range(len(records)))11result = []⋯values this step[0, 1, 2, 3]remaining
9result ← []⋯10remaining = list(range(len(records)))11result = []12for _ in range(len(records)):⋯values this step[]result
10_ ← 0⋯11result = []12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])⋯values this step0_
11best ← 1⋯12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])14    result.append(names[best])⋯values this step1best
12result ← ['bo']⋯13best = min(remaining, key=lambda i: scores[i])14result.append(names[best])15remaining.remove(best)⋯values this step[] → ['bo']result
13remaining ← [0, 2, 3]⋯14    result.append(names[best])15    remaining.remove(best)16print('RESULT:', result)values this step[0, 1, 2, 3] → [0, 2, 3]remaining
14_ ← 1⋯11result = []12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])⋯values this step0 → 1_
15best ← 3⋯12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])14    result.append(names[best])⋯values this step1 → 3best
16result ← ['bo', 'di']⋯13best = min(remaining, key=lambda i: scores[i])14result.append(names[best])15remaining.remove(best)⋯values this step['bo'] → ['bo', 'di']result
17remaining ← [0, 2]⋯14    result.append(names[best])15    remaining.remove(best)16print('RESULT:', result)values this step[0, 2, 3] → [0, 2]remaining
18_ ← 2⋯11result = []12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])⋯values this step1 → 2_
19best ← 0⋯12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])14    result.append(names[best])⋯values this step3 → 0best
20result ← ['bo', 'di', 'al']⋯13best = min(remaining, key=lambda i: scores[i])14result.append(names[best])15remaining.remove(best)⋯values this step['bo', 'di'] → ['bo', 'di', 'al']result
21remaining ← [2]⋯14    result.append(names[best])15    remaining.remove(best)16print('RESULT:', result)values this step[0, 2] → [2]remaining
22_ ← 3⋯11result = []12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])⋯values this step2 → 3_
23best ← 2⋯12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])14    result.append(names[best])⋯values this step0 → 2best
24result ← ['bo', 'di', 'al', 'cy']⋯13best = min(remaining, key=lambda i: scores[i])14result.append(names[best])15remaining.remove(best)⋯values this step['bo', 'di', 'al'] → ['bo', 'di', 'al', 'cy']result
25remaining ← []⋯14    result.append(names[best])15    remaining.remove(best)16print('RESULT:', result)values this step[2] → []remaining
26for _ in range(len(records)):⋯11result = []12for _ in range(len(records)):13    best = min(remaining, key=lambda i: scores[i])⋯
27stdout ← RESULT: ['bo', 'di', 'al', 'cy']⋯15    remaining.remove(best)16print('RESULT:', result)values this stepRESULT: ['bo', 'di', 'al', 'cy']stdout

With pandas

df.sort_values('score') returns a new sorted DataFrame — the original df is unchanged. The snapshot shows names, scores, and the preserved (non-reset) original index.

                            library.py
                        

                            import pandas as pd
from dalib.display import set_display
set_display()

records = [
    {'name': 'al', 'score': 80},
    {'name': 'bo', 'score': 55},
    {'name': 'cy', 'score': 90},
    {'name': 'di', 'score': 70},
]
df = pd.DataFrame(records)
sorted_df = df.sort_values('score')
print('names:', sorted_df['name'].tolist())
print('scores:', sorted_df['score'].tolist())
print('index:', sorted_df.index.tolist())
print('RESULT:', sorted_df['name'].tolist())

                        

names: ['bo', 'di', 'al', 'cy']
scores: [55, 70, 80, 90]
index: [1, 3, 0, 2]
RESULT: ['bo', 'di', 'al', 'cy']

Implementation notes

sort_values returns a new DataFrame; the original is unchanged. Use inplace=True to sort in place, but prefer the default for readable pipelines.
The sorted result keeps the original row index ([1, 3, 0, 2] above). Add .reset_index(drop=True) to renumber from 0.
Default order is ascending. Pass ascending=False for descending: df.sort_values('score', ascending=False).
Cross-reference: sort-by-key (python-data-basics) for the pure-Python version; sort-array (python-numpy ch08) for the NumPy array version.