What problem does it solve?
MATLAB data I/O involves subtle pitfalls: European locale delimiters, silent data corruption (Excel Inf becoming 65535, Parquet integer promotion), misleading error messages, and legacy patterns like fopen/fgetl loops. This Skill provides correct, modern patterns for importing and exporting tabular and structured data while avoiding these traps.
Core Features & Use Cases
- Format-specific guidance: Correct usage of readtable/writetable, detectImportOptions, readtimetable, readstruct/writestruct, and readdictionary for CSV, Excel, Parquet, JSON, and XML files.
- Validation and error diagnosis: Post-import checks for silent corruption (NaN columns, 65535 values, int64-to-double promotion) and interpretation of misleading error messages.
- Locale and URL handling: European delimiter/decimal-separator configuration, reading from URLs and authenticated REST endpoints, and direct reads from compressed archives.
- Use Case: Import a German semicolon-delimited CSV with comma decimal separators, validate the imported table for missing values, and export the cleaned results to Parquet.
Quick Start
Ask your agent to import a CSV or Excel file into MATLAB using readtable with proper import options and validate the results.