No column indexing - cannot reference columns by position when names have special characters

## Description

xl does not support referencing columns by index/position (e.g., column 1, 2, 3). This is critical when column names contain special characters that break filter/group commands (see XL-001).

## Current Behavior

```bash
# Column name "Qté VTE" has special character "é"
# Cannot filter with name:
$ xl filter "sales.xlsx" -s DONNEES "`Qté VTE` < 0"
/usr/bin/bash: line 1: Qté: command not found

# Cannot reference by index either:
$ xl filter "sales.xlsx" -s DONNEES -c 7 "< 0"
Error: Column "7" not found

# No workaround available in xl
```

## Expected Behavior

Users should be able to reference columns by position:

```bash
# Reference column 7 by index
$ xl filter "sales.xlsx" -s DONNEES --col 7 "< 0"

# Or with -c shorthand
$ xl filter "sales.xlsx" -s DONNEES -c 7 "< 0"

# Or with @ notation
$ xl filter "sales.xlsx" -s DONNEES "@7 < 0"
```

## Steps to Reproduce

1. Create Excel file with column named "Qté VTE" (contains accent)
2. Try to filter on column by name: Fails (XL-001)
3. Try to filter on column by index: Fails (this issue)
4. No way to filter on this column

## Real-World Impact

### Use Case: Filter on Numeric Column with Special Characters

**Business Question:** Find outlier transactions (negative quantities, bulk orders)

**Column:** `Qté VTE` (Quantity Sold in French)
**Position:** Column 7

**Current Approach (Impossible):**
```bash
# Cannot use name (special character bug)
xl filter sales.xlsx -s DONNEES "`Qté VTE` < 0"
# Error: bash interpretation error

# Cannot use index (not supported)
xl filter sales.xlsx -s DONNEES --col 7 "< 0"
# Error: Column "7" not found

# Must export to Python
python -c "import pandas as pd; df = pd.read_excel('sales.xlsx'); print(df[df.iloc[:, 6] < 0])"
```

**Proposed Approach (with column indexing):**
```bash
xl filter sales.xlsx -s DONNEES --col 7 "< 0"
# Works immediately
```

### Use Case: Dynamic Column Reference

**Business Question:** Filter on last column without knowing name

**Current:** Must inspect file first, get column name
**Proposed:** `xl filter file.xlsx --col -1 "< 100"` (last column)

### Use Case: Anonymous Data

**Business Question:** Analyze CSV/file without headers

**Current:** Cannot reference columns easily
**Proposed:** `xl group file.csv --by @1 --aggregate "@3:sum"`

## Proposed Solution

Add column index/position reference support to all column-accepting commands.

### Syntax Options

**Option 1: --col INDEX (Preferred)**
```bash
xl filter file.xlsx --col 7 "condition"
```

**Option 2: -c INDEX (Shorthand)**
```bash
xl filter file.xlsx -c 7 "condition"
```

**Option 3: @INDEX notation**
```bash
xl filter file.xlsx "@7 < 0"
```

**Option 4: Negative indexing (Python style)**
```bash
xl filter file.xlsx --col -1 "< 100"  # Last column
xl filter file.xlsx --col -2 "< 50"   # Second-to-last column
```

**Recommendation:** Support Options 1, 2, and 4 for maximum flexibility

## Affected Commands

All commands that accept column names should also support column indices:

| Command | Column Flag | Should Support Index? |
|---------|-------------|----------------------|
| `xl filter` | `--condition` (implicit) | ✅ Yes |
| `xl group` | `--by` | ✅ Yes |
| `xl unique` | `-c, --columns` | ✅ Yes |
| `xl count` | `-c, --columns` | ✅ Yes |
| `xl sort` | `--by` | ✅ Yes |
| `xl select` | `-c, --columns` | ✅ Yes |
| `xl stats` | `-c, --columns` | ✅ Yes |

## Acceptance Criteria

- [ ] Can filter on column by positive index (1, 2, 3, ...)
- [ ] Can filter on column by negative index (-1, -2, -3 for last, 2nd last, ...)
- [ ] Can group by column index
- [ ] Can sort by column index
- [ ] Can get unique values by column index
- [ ] Can get stats by column index
- [ ] Works with Excel files
- [ ] Works with CSV files
- [ ] Documentation updated with examples
- [ ] Error message shows: "Did you mean column index? Use --col N"

## Test Cases

```bash
# Test 1: Filter by positive index
xl filter sales.xlsx -s DONNEES --col 7 "< 0"
# Expected: Rows where column 7 < 0

# Test 2: Filter by negative index (last column)
xl filter sales.xlsx --col -1 "> 100"
# Expected: Rows where last column > 100

# Test 3: Group by column index
xl group sales.xlsx --by 2 --aggregate 5:sum
# Expected: Group by column 2, sum column 5

# Test 4: Sort by column index
xl sort sales.xlsx --by 3 --order desc
# Expected: Sort by column 3 descending

# Test 5: Stats by column index
xl stats sales.xlsx -c 4
# Expected: Statistics for column 4

# Test 6: Unique by column index
xl unique sales.xlsx -c 2
# Expected: Unique values in column 2

# Test 7: Mixed names and indices
xl group sales.xlsx --by "Circuit,3" --aggregate 5:sum
# Expected: Group by Circuit name and column 3

# Test 8: Aggregation by index
xl group sales.xlsx --by 2 --aggregate "5:sum,6:mean"
# Expected: Group by col 2, sum col 5, mean col 6
```

## Implementation Notes

### Index Convention

**Important:** 1-based indexing (user-friendly) vs 0-based (Python-style)?

**Recommendation:** 1-based for CLI users (more intuitive):
- Column 1 = First column
- Column -1 = Last column
- Matches Excel/SQL conventions

**Alternative:** Support both with clear documentation

### Detection Logic

```python
# Pseudo-code
def resolve_column_reference(col_ref, df):
    if isinstance(col_ref, int):
        # Column index
        return df.columns[col_ref - 1]  # Convert 1-based to 0-based
    elif col_ref.startswith('@'):
        # @ notation
        idx = int(col_ref[1:])
        return df.columns[idx - 1]
    elif col_ref.lstrip('-').isdigit():
        # Negative index
        return df.columns[int(col_ref)]
    else:
        # Column name (existing behavior)
        return col_ref
```

### Error Messages

```bash
# When column name not found
$ xl filter sales.xlsx "NonExistent < 0"
Error: Column 'NonExistent' not found
Available columns: DateVente, jour, Circuit, Magasin, Catégorie, Libellé, Qté VTE
Hint: Use column index (e.g., --col 7) for special characters

# When index out of range
$ xl filter sales.xlsx --col 20 "< 0"
Error: Column index 20 out of range (file has 7 columns)
```

## Related Issues

- **XL-001:** Special characters in filter (primary use case for column indexing)
- **XL-011:** Inconsistent option naming (this adds consistency)
- **XL-003:** No sorting in group (could use index for sort column)

## Workarounds

### Current Workaround 1: Rename columns first

```bash
# If xl rename worked (future feature)
xl rename sales.xlsx --from "Qté VTE" --to "Qty"
xl filter sales.xlsx "Qty < 0"
```

### Current Workaround 2: Export to Python/pandas

```python
import pandas as pd
df = pd.read_excel('sales.xlsx')
# Can use iloc for position-based indexing
result = df[df.iloc[:, 6] < 0]  # Column 7 (0-based = 6)
```

### Current Workaround 3: Use Excel

```bash
# Open in Excel
# Filter by column position (column G)
# Save filtered file
```

## Benefits

1. **Special Character Workaround:** Fixes XL-001 without renaming
2. **Anonymous Data:** Analyze files without headers
3. **Dynamic Scripts:** Reference columns without knowing names
4. **International Users:** Work around character encoding issues
5. **Quick Analysis:** No need to memorize column names
6. **Robustness:** Works even when column names change

## Industry Standards

### pandas
```python
df.iloc[:, 6]  # Column 7 (0-based indexing)
```

### SQL
```sql
-- No standard column indexing, but some databases support:
SELECT * FROM table ORDER BY 3  # Order by 3rd column
```

### Excel
```
Column G = 7th column (1-based)
```

**Recommendation:** Follow Excel convention (1-based indexing)

## Advanced Features (Future)

### Range Notation
```bash
xl select sales.xlsx --columns "1-5"  # Select columns 1 through 5
```

### Multiple Indices
```bash
xl group sales.xlsx --by "1,2,3" --aggregate "5:sum"
```

### Header Detection
```bash
xl filter sales.csv --no-header --col 3 "< 0"
# For CSV files without header row
```

## Labels

`feature-request` `high-priority` `workaround` `column-indexing` `special-characters` `usability`

Command	Column Flag	Should Support Index?
`xl filter`	`--condition` (implicit)	✅ Yes
`xl group`	`--by`	✅ Yes
`xl unique`	`-c, --columns`	✅ Yes
`xl count`	`-c, --columns`	✅ Yes
`xl sort`	`--by`	✅ Yes
`xl select`	`-c, --columns`	✅ Yes
`xl stats`	`-c, --columns`	✅ Yes

No column indexing - cannot reference columns by position when names have special characters #12

Description

Description

Current Behavior

Expected Behavior

Steps to Reproduce

Real-World Impact

Use Case: Filter on Numeric Column with Special Characters

Use Case: Dynamic Column Reference

Use Case: Anonymous Data

Proposed Solution

Syntax Options

Affected Commands

Acceptance Criteria

Test Cases

Implementation Notes

Index Convention

Detection Logic

Error Messages

Related Issues

Workarounds

Current Workaround 1: Rename columns first

Current Workaround 2: Export to Python/pandas

Current Workaround 3: Use Excel

Benefits

Industry Standards

pandas

SQL

Excel

Advanced Features (Future)

Range Notation

Multiple Indices

Header Detection

Labels

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions