Skip to main content

Import & export

CSV, TSV, HTML, XML, Markdown, JSON, XLSX, and PDF.

Import​

sm.importCsv(text);  sm.importTsv(text);  sm.importHtml(text);
sm.importMarkdown(text);  sm.importXml(text);
sm.importJson((jsonDecode(text) as List).cast<Map<String, dynamic>>());
sm.importXlsx(bytes);                    // Uint8List

Transform before import (Power-Query lite)​

Reshape a JSON payload to the grid's columns with a chainable, non-mutating pipeline, then importJson the result:

final rows = BCellImportTransform(raw)
    .rename({'full_name': 'name'})       // key -> new key
    .select(['name', 'age'])             // keep only these
    .where((r) => (r['age'] as num) >= 18)
    .mapColumn('name', (v) => '$v'.trim())
    .distinct()                          // drop duplicate rows
    .rows;                               // also: drop / addColumn / sortBy
sm.importJson(rows);

No joins / grouping / aggregation — heavy transforms belong in the host data layer.

Export​

sm.toCsv();  sm.toTsv();  sm.toHtml();  sm.toXml();  sm.toMarkdown();
sm.toJson();                             // List<Map> ready for jsonEncode
sm.toXlsx();                             // -> List<int> (.xlsx bytes)
sm.toXlsx(formulas: true);               // export '=A1+B1' as a real Excel formula
sm.toXlsx(styled: false, merged: false); // plain values only (styles/merges are on)
await sm.toPdf(title: 'People');         // -> PDF bytes
await sm.toPdf(title: 'People', selectionOnly: true); // print area = current range

XLSX carries fidelity by default: bold header, per-column alignment and wrap-text, and cellStyle fill/font colour (styled), plus colspan/rowspan as real Excel merges that value-round-trip through importXlsx (merged). Column widths ride both directions (Excel width units ≈ px/7, clamped to minWidth on import). With formulas: true, =... cells export as real Excel formulas and importXlsx maps them back to live =... strings, so formulas survive the round-trip. Number-format codes don't ride — the excel package's writer can't emit custom numFmts.

Every other export emits the evaluated value of formula cells (never the raw = — the XLSX default too), and rich-text/number-format cells flatten to plain text.