Source: postgresql-9.2.1\src\backend\utils\adt\json.c:
/*
* Input.
*/
Datum
json_in(PG_FUNCTION_ARGS)
{
char *text = PG_GETARG_CSTRING(0);
json_validate_cstring(text);
/* Internal representation is the same as text, for now */
PG_RETURN_TEXT_P(cstring_to_text(text));
}
Update for PostgreSQL 9.3.5:
The code has changed in the json_in function, but the json internal representation is still text:
Source: postgresql-9.3.5\src\backend\utils\adt\json.c:
/*
* Input.
*/
Datum
json_in(PG_FUNCTION_ARGS)
{
char *json = PG_GETARG_CSTRING(0);
text *result = cstring_to_text(json);
JsonLexContext *lex;
/* validate it */
lex = makeJsonLexContext(result, false);
pg_parse_json(lex, &nullSemAction);
/* Internal representation is the same as text, for now */
PG_RETURN_TEXT_P(result);
}
So it appears that, for now at least, json is the same as a text datatype but with JSON validation. The text datatype's maximum size is 1GB.
For jsonb fields, if you look at the jsonb.c in the source code, you'll see this function:
checkStringLen(size_t len)
{
if (len > JENTRY_OFFLENMASK)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("string too long to represent as jsonb string"),
errdetail("Due to an implementation restriction, jsonb strings cannot exceed %d bytes.",
JENTRY_OFFLENMASK)));
return len;
}
Error code says that jsonb strings cannot exceed JENTRY_OFFLENMASK bytes.