In-memory output - Jetxl

The _to_bytes variants take the same arguments as their file counterparts, minus filename, and hand back bytes.

excel_bytes = jet.write_sheet_arrow_to_bytes(
    df.to_arrow(),
    sheet_name="Employees",
    styled_headers=True,
    auto_width=True,
)

This matters wherever the filesystem is awkward: read-only containers, Lambda’s ephemeral storage, or any web server where writing a temp file means cleaning it up afterwards.

Serving from a web framework

from fastapi import FastAPI
from fastapi.responses import Response

XLSX = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"

app = FastAPI()

@app.get("/report")
async def report():
    data = jet.write_sheet_arrow_to_bytes(df.to_arrow(), styled_headers=True)
    return Response(
        content=data,
        media_type=XLSX,
        headers={"Content-Disposition": "attachment; filename=report.xlsx"},
    )

from flask import Flask, Response

XLSX = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"

app = Flask(__name__)

@app.route("/download")
def download():
    data = jet.write_sheet_arrow_to_bytes(df.to_arrow(), styled_headers=True)
    return Response(
        data,
        mimetype=XLSX,
        headers={"Content-Disposition": "attachment;filename=data.xlsx"},
    )

import base64

XLSX = "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"

def lambda_handler(event, context):
    data = jet.write_sheet_arrow_to_bytes(df.to_arrow())
    return {
        "statusCode": 200,
        "body": base64.b64encode(data).decode("utf-8"),
        "isBase64Encoded": True,
        "headers": {"Content-Type": XLSX},
    }

That long media type is the correct one for .xlsx. Serving application/vnd.ms-excel instead, which is the older .xls type, makes some browsers mislabel the download.

Straight to object storage

import boto3

s3 = boto3.client("s3")
s3.put_object(
    Bucket="my-bucket",
    Key="reports/monthly.xlsx",
    Body=jet.write_sheets_arrow_to_bytes(sheets, num_threads=2),
    ContentType=(
        "application/vnd.openxmlformats-officedocument."
        "spreadsheetml.sheet"
    ),
)

Multiple sheets to bytes

excel_bytes = jet.write_sheets_arrow_to_bytes(
    [
        {"data": df1.to_arrow(), "name": "Sales", "styled_headers": True},
        {"data": df2.to_arrow(), "name": "Costs", "auto_width": True},
    ],
    num_threads=2,
)

Here num_threads defaults to 1, unlike the file-writing write_sheets_arrow where it’s required.

Memory

Jetxl holds the entire workbook in memory before returning it. For a million-row sheet that’s on the order of a gigabyte during generation, which is fine on a server and potentially fatal in a small Lambda. Write to a file for very large exports, or split the work.