I/O API

Methods for reading data into and writing data out of Molniya.

Reading Data

readCsv

Reads a CSV file into a DataFrame.

readCsv(path: string, options?: CsvReadOptions): Promise<{ df: DataFrame, errors: CsvError[] }>

Returns: An object containing:

• df: The resulting DataFrame.
• errors: An array of parsing errors (if any were strictly malformed but recoverable).

Options:

interface CsvReadOptions {
  delimiter?: string;       // Default: ','
  hasHeader?: boolean;      // Default: true
  encoding?: string;        // Default: 'utf-8'
  maxRows?: number;         // Stop after N rows
  sampleRows?: number;      // Rows to scan for type inference (Default: 100)
  nullValues?: string[];    // Strings to treat as null (e.g. ['NA', 'null'])
  datetime?: {
    defaultZone?: string;   // Default timezone (e.g. 'UTC')
    columns?: Record<string, { format: 'iso'|'sql'|'unix-s'|'unix-ms', zone?: string }>
  }
}

Example:

const { df } = await readCsv('./data.csv', {
  delimiter: '\t',
  maxRows: 500
});

---

scanCsv (Lazy Loading)

Creates a LazyFrame for processing files larger than memory.

scanCsv(path: string, options?: CsvReadOptions): Promise<LazyFrame>

NOTE

LazyFrame supports a subset of operations (filter, select, head) and executes them only when you call .collect().

---

Writing Data

writeCsv

Writes a DataFrame to a CSV file.

writeCsv(df: DataFrame, path: string): Promise<void>

Example:

await writeCsv(resultDf, './report.csv');

---

toCsv

Converts a DataFrame to a CSV string.

toCsv(df: DataFrame): string

Example:

const csvString = toCsv(df);
console.log(csvString);