@kanunilabs/pivotgrid-core 1.6.0 → 1.7.1

package/README.md

@@ -31,13 +31,13 @@
 
 ## Engine highlights
 
-- **Web Worker computation** — all aggregation runs off the main thread; the UI never blocks.
-- **Columnar storage** — dictionary-encoded `Int32Array` / `Float64Array` columns for memory efficiency on multi-million-row datasets.
+- **Web Worker computation** — grouping and the built-in aggregations run off the main thread. A `custom` summary is your own function, so it runs on the main thread over rows the worker grouped.
+- **Columnar storage** — dictionary-encoded `Int32Array` / `Float64Array` columns for memory efficiency on million-row datasets.
 - **FastTrie grouping** — `Int32Array`-based trie for hierarchical row/column grouping.
 - **Pipelined incremental init** — data is encoded and streamed to the worker in chunks, so encoding and aggregation overlap instead of running back-to-back.
 - **Reconfigure fast path** — moving a field re-groups the data already held in the worker; no re-encode, no re-transfer.
 - **Bitmask filtering** — O(n) `Uint8Array` row masks applied inside the worker.
-- **Summary display modes** — % of grand total / row / column, running total, rank, index and more.
+- **Summary display modes** — % of grand total / row / column, running total, difference and % difference.
 - **CSP-safe formulas** — recursive-descent expression parser; no `eval`, no `new Function`.
 - **Import & export** — CSV / JSON / XML export and file/clipboard import with type detection.
 - **12 locales** — en, tr, de, fr, es, it, pt, ru, ar, zh-CN, ja, ko (RTL included).