Data · SQL
A 50 MB SQL file, on demand.
Exactly 52,428,800 bytes of a runnable dump with a CREATE TABLE and an INSERT, built on your machine the moment you press the button.
52,428,800 bytes, exactly.
Every byte above the 126 bytes structural minimum goes into a block comment, which is why this file still behaves like one a real user would send.
Unit convention
50 MB on this page means exactly 52,428,800 bytes, using the binary (MiB) convention some systems label "MB" for short. 1 MB in the strict decimal sense is 1,000,000 bytes; 1 MiB is 1,048,576 bytes — the readout above always shows the exact count either way.
Approximate size equivalents
Roughly 29,127 pages of typed text, 55 minutes of 128 kbps audio, 5 minutes of CD-quality audio, 21 photos from a 12 MP camera. These figures compare file size only — the generated file does not contain this text, audio or image content.
Accepted by
Cloudflare Free proxy body, GitHub repository file, Slack free workspace file — all publish a limit at or above this size, based on commonly cited figures that can vary by plan and change over time.
Rejected by
Gmail attachment, Outlook.com attachment, Discord free attachment — this exceeds their commonly cited limit, which makes it a useful negative test. Published limits can vary by plan and change over time.
Other SQL sizes
Files this large exercise multipart and chunked upload paths rather than a single request, which is usually where the bugs are.
Questions.
Is this 50 MB SQL file valid?
Yes. It is a runnable dump with a CREATE TABLE and an INSERT. The extra bytes sit in a block comment, so it opens in psql, mysql, DBeaver and TablePlus.
Does 50 MB mean 52,428,800 bytes?
On this page, yes — 50 MB uses the binary (MiB) convention, so the file is exactly 52,428,800 bytes. 1 MB in the strict decimal sense would be a different, smaller byte count. Upload validators may use either convention, so the exact byte count is always shown to remove any ambiguity.
What is a 50 MB file used for?
Files this large exercise multipart and chunked upload paths rather than a single request, which is usually where the bugs are. For SQL specifically, the usual job is testing database import limits and migration tooling.
Where is the file downloaded from?
Nowhere. It is assembled in a Web Worker on your own device and handed to the browser as a Blob, so nothing is uploaded, fetched or stored.