Convert HTML Character Entities to UTF-8

The conversion

// PHP — name the charset explicitly
html_entity_decode($s, ENT_QUOTES | ENT_HTML5, 'UTF-8');

# Python
import html
html.unescape('café')      # 'café'

// JavaScript — let the parser do it
new DOMParser().parseFromString(s, 'text/html').documentElement.textContent;

All three handle named, decimal and hexadecimal references. Hand-written replacement chains cover four entities and miss two thousand.

Why the charset argument matters

The decoder produces characters and must then encode them as bytes. Ask PHP for ISO-8859-1 output and → (→) has no representation, so it is dropped or replaced.

Specify UTF-8 for both the input interpretation and the output, and make sure the destination — file, database column, HTTP header — is also UTF-8. A correct decode written into a Latin-1 column corrupts on write.

Cleaning up stored content

Entity-encoded text in a database is a liability: JSON APIs, CSV exports and emails all show the raw é. Decode once in a migration rather than at every render.

Watch for double-encoded values — é decodes to é and needs a second pass. Handle those deliberately as a one-off data fix with a manual review, never with a loop in production code.

Confirm the column and connection are UTF-8 first (utf8mb4 on MySQL, not utf8), or the migration will replace one corruption with another.