Streams API (High Performance I/O)

For very large datasets, pyopenxlsx provides streaming I/O that bypasses dense Cell object graphs. Worksheet.stream_writer() / stream_reader() return high-level StreamWriter / StreamReader façades over the native stream types.

Important: While a stream writer is active on a worksheet, avoid mixing standard DOM cell writes until the writer is closed.

Stream Writer

StreamWriter appends rows sequentially. Date/datetime values use the same coercion and Workbook.auto_date_formats behaviour as bulk worksheet APIs (values are written as Excel serials, with optional style indices).

from datetime import date
from pyopenxlsx import Workbook, Font

with Workbook() as wb:
    ws = wb.active
    bold = wb.add_style(font=Font(bold=True))

    with ws.stream_writer() as writer:
        writer.append_row(["ID", "When", "Name", "Score"])
        writer.append_row([(1, bold), date(2023, 1, 1), ("Alice", bold), 99.9])
        for i in range(1_000_000):
            writer.append_row([i, date(2023, 1, 1), f"User_{i}", 99.9])

    wb.save("large_output.xlsx")

Explicit (value, style_index) tuples are preserved. Plain values inherit default formatting (or auto date styles when enabled).

Methods

append_row(values, row_opts=None)
set_row(row, start_col, values, row_opts=None) / set_row_ref(ref, values, ...)
Context manager (with writer:) and close()
Properties: is_active, last_row, max_column

Stream Reader

Iterate rows without loading the full worksheet DOM into Python Cell wrappers.

from pyopenxlsx import Workbook

with Workbook("large_input.xlsx") as wb:
    ws = wb.active
    reader = ws.stream_reader()
    for row_data in reader:
        idx = reader.current_row_index
        # process row_data (list of values)

Optional XLStreamReadOptions (or kwargs empty_rows / apply_number_formats) control empty-row policy and number-format application.

Use cases

Exporting database query results directly to Excel
Parsing multi-gigabyte workbooks where a full DOM would OOM