prize18-8  Exercise 1: Characters Against Bytes in SQL
======================================================

Plan: length(value) gives code points; length(CAST(value AS BLOB)) gives bytes. ASCII passes both limits only when it is at most 5 long. Anything with
multi-byte characters can pass the character limit and still fail the byte limit.

ex1.py:

  import sqlite3

  con = sqlite3.connect(":memory:")
  values = ["hello", "caf\u00e9 bar", "\u20ac\u20ac\u20ac\u20ac", "\U0001F600" * 3, "\u8a9e\u8a9e\u8a9e"]
  for v in values:
      chars, nbytes = con.execute("SELECT length(?), length(CAST(? AS BLOB))", (v, v)).fetchone()
      print(f"{ascii(v):28} chars={chars:2} bytes={nbytes:2}  fits 5 chars: {chars <= 5!s:5} fits 5 bytes: {nbytes <= 5}")

Real output (Python 3.14, bundled SQLite):

  'hello'                      chars= 5 bytes= 5  fits 5 chars: True  fits 5 bytes: True
  'caf\xe9 bar'                chars= 8 bytes= 9  fits 5 chars: False fits 5 bytes: False
  '\u20ac\u20ac\u20ac\u20ac'   chars= 4 bytes=12  fits 5 chars: True  fits 5 bytes: False
  '\U0001f600\U0001f600\U0001f600' chars= 3 bytes=12  fits 5 chars: True  fits 5 bytes: False
  '\u8a9e\u8a9e\u8a9e'         chars= 3 bytes= 9  fits 5 chars: True  fits 5 bytes: False

The euro, emoji and CJK strings pass a five-character limit but would fail a five-byte limit. If a product measures a column or an index key in bytes
(check for yours), these values would be rejected or truncated even though they look short, and the problem appears only for non-ASCII data, which makes
it easy to miss in testing with English sample text.