Out-of-order notebook, caught and flagged

Execution-order warnings and stale-state detection on a messy notebook.

Fidelity warnings

• Cell [2]: execution order changed (In[2] appears after In[5]) — outputs may not reflect top-to-bottom state
• `totals` used in Cell [2] but defined only in Cell [3] — value likely came from an earlier run
• 1 long output(s) truncated to 30 lines

Converted Markdown

# out-of-order

## Notebook overview

- Source: `out-of-order.ipynb` · Python 3 (python)
- Cells: 5 (5 code · 0 markdown · 1 not executed)
- Execution: ⚠ 1 cell(s) executed out of notebook order
- Imports: `pd` — imported names, excluded from dependency hints

> Dependency annotations ("depends on") are approximate cell dependency hints from static regex analysis of assignments and imports — not full dataflow analysis. Re-assignments inside branches, dynamic scope tricks and string-built code can fool them.

## Cell [3] · type:code · id:f0f0f0f0

```python
import pandas as pd
totals = pd.Series([1, 2, 3])
```

## Cell [5] · type:code · id:e1e1e1e1

> ⚠️ depends on: `totals` (defined in Cell [3])

```python
summary = totals.describe()
print(summary)
```

**Output:**

```
row 0: value=0
row 1: value=3
row 2: value=6
row 3: value=9
row 4: value=12
row 5: value=15
row 6: value=18
row 7: value=21
row 8: value=24
row 9: value=27
row 10: value=30
row 11: value=33
row 12: value=36
row 13: value=39
row 14: value=42
row 15: value=45
row 16: value=48
row 17: value=51
row 18: value=54
row 19: value=57
row 20: value=60
row 21: value=63
row 22: value=66
row 23: value=69
row 24: value=72
row 25: value=75
row 26: value=78
row 27: value=81
row 28: value=84
row 29: value=87
… [output truncated: 170 more lines]
```

## Cell [2] · type:code · id:d2d2d2d2

> ⚠ execution order changed — In[2] appears after In[5] in notebook order

```python
print(len(totals))
print("cleaned")
```

**Output:**

````
```
not a real fence
```
````

## Cell [8] · type:code · id:c3c3c3c3

> ⚠️ depends on: `summary` (defined in Cell [5])

```python
report = summary.to_string()
print(report[:40])
```

**Output:**

```
count 3.0
```

## Cell [p5] · type:code (not executed) · id:b4b4b4b4

```python
scratch = 1
```

Keywords: jupyter execution order, notebook reproducibility llm · All examples