Context Slice Recognition
- "spreadsheet", "data table", "csv", "export data"
- Numerical data, records, tabular information
- Data for analysis, import, or archival
Structure
Header1,Header2,Header3,Header4
value1,value2,value3,value4
value1,value2,value3,value4
First row is always headers. Use clear, descriptive column names.
Rules
| Element | Format |
|---|---|
| Headers | Title Case, no spaces (use underscores: First_Name) |
| Text with commas | Wrap in quotes: "Smith, John" |
| Currency | No symbol, 2 decimals: 1234.56 (not $1,234.56) |
| Percentages | Decimal form: 0.15 (not 15%) |
| Dates | ISO 8601: 2024-12-04 |
| Empty values | Leave blank, don't use "N/A" or "null" |
| Boolean | true / false (lowercase) |
Column Ordering
Order columns by importance and logical grouping:
- Identifiers first (ID, name, date)
- Primary data next (the main values)
- Calculated/derived after (totals, percentages)
- Metadata last (notes, sources, timestamps)
Financial Data Pattern
Date,Description,Category,Amount,Running_Balance
2024-12-01,Starting Balance,Opening,0,10000.00
2024-12-02,Client Payment,Revenue,5000.00,15000.00
2024-12-03,Software License,Expense,-299.00,14701.00
Data Export Pattern
ID,Name,Email,Created_Date,Status,Last_Activity
001,John Smith,john@example.com,2024-01-15,active,2024-12-01
002,Jane Doe,jane@example.com,2024-03-22,inactive,2024-06-15
Pitfalls
- Including currency symbols (breaks numeric parsing)
- Inconsistent date formats (always ISO 8601)
- Spaces in headers (use underscores)
- Missing quotes around text with commas
- Including formulas (CSV is data only, no formulas)