TO_BINARY_FLOAT
TO_BINARY_FLOAT converts its argument to a single-precision BINARY_FLOAT. Use TO_BINARY_FLOAT(value) for numeric or binary floating-point inputs, or for text in a form such as '1.25' or '1.25e6'. Use TO_BINARY_FLOAT(text, format) when a string contains delimiters described by a numeric format model.
SELECT TO_BINARY_FLOAT('1,234.5', '9,999.9') AS formatted,
TO_BINARY_FLOAT(16777217::NUMBER) =
TO_BINARY_FLOAT(16777216::NUMBER) AS rounded,
TO_BINARY_FLOAT(NULL::TEXT) IS NULL AS null_preserved
FROM dual;
The query returns 1234.5, t, and t. Single precision rounds 16777217 to 16777216. When that distinction matters, see TO_BINARY_DOUBLE.