importer

type FuzzyImportOptions struct {
	*ImportOptions

	// SkipInvalidRows skips rows that don't match the expected column count (default true)
	SkipInvalidRows bool

	// TrimWhitespace aggressively trims whitespace from all values (default true)
	TrimWhitespace bool

	// FixQuotes attempts to fix unmatched quotes in CSV data (default true)
	FixQuotes bool

	// CoerceTypes attempts to coerce invalid values to the expected type (default true)
	CoerceTypes bool

	// MaxSkippedRows maximum number of rows to skip before giving up (default 100)
	MaxSkippedRows int

	// AllowMixedTypes allows columns to have mixed types (converts to TEXT) (default true)
	AllowMixedTypes bool

	// FuzzyJSON attempts to parse malformed JSON (default true)
	FuzzyJSON bool

	// RemoveInvalidChars removes invalid UTF-8 characters (default true)
	RemoveInvalidChars bool

	// AutoFixDelimiters tries to detect and fix inconsistent delimiters (default true)
	AutoFixDelimiters bool
}

type ImportOptions struct {
	// BatchSize controls how many rows are buffered before executing INSERT (default 1000).
	BatchSize int

	// NullLiterals are treated as SQL NULL (case-insensitive, trimmed).
	// Defaults: "", "null", "na", "n/a", "none", "#n/a"
	NullLiterals []string

	// CreateTable creates the table if it doesn't exist using inferred column types (default true).
	// Disable this if you want to create the table manually with specific types/constraints.
	CreateTable bool

	// Truncate clears the table before import (default false).
	Truncate bool

	// HeaderMode controls header detection:
	//   "auto" (default)  → heuristic decides based on data analysis
	//   "present"         → first row is always treated as header
	//   "absent"          → first row is data, synthetic column names generated (col_1, col_2, ...)
	HeaderMode string

	// DelimiterCandidates tested during auto-detection. Default: , ; \t |
	// Override to force specific delimiter(s).
	DelimiterCandidates []rune

	// TableName overrides the target table name (useful when importing from stdin/pipes).
	TableName string

	// SampleBytes caps the amount of data used for detection (default 128KB).
	SampleBytes int

	// SampleRecords caps the number of records analyzed for detection (default 500).
	SampleRecords int

	// TypeInference controls automatic type detection (default true).
	// When enabled, analyzes sample data to determine INT, FLOAT, BOOL, TIME, or TEXT.
	// When disabled, all columns default to TEXT type.
	TypeInference bool

	// DateTimeFormats lists custom datetime formats to try during type detection.
	// Defaults include RFC3339, ISO8601, common US/EU formats.
	DateTimeFormats []string

	// StrictTypes when true causes import to fail if data doesn't match detected types (default false).
	// When false, falls back to TEXT on type conversion errors.
	StrictTypes bool
}

type ImportResult struct {
	RowsInserted int64             // Total rows successfully inserted
	RowsSkipped  int64             // Rows skipped due to errors (if StrictTypes=false)
	Delimiter    rune              // Detected or configured delimiter
	HadHeader    bool              // Whether a header row was detected/configured
	Encoding     string            // Detected encoding: "utf-8", "utf-8-bom", "utf-16le", "utf-16be"
	ColumnNames  []string          // Final column names used
	ColumnTypes  []storage.ColType // Detected column types
	Errors       []string          // Non-fatal errors encountered during import
}

type XMLRecord struct {
	XMLName xml.Name
	Attrs   []xml.Attr  `xml:",any,attr"`
	Content string      `xml:",chardata"`
	Nodes   []XMLRecord `xml:",any"`
}