The doc for "quote_ident()" says this:
«
https://www.postgresql.org/docs/14/functions-string.html
Returns the given string suitably quoted to be used as an identifier in an SQL
statement string. Quotes are added only if necessary (i.e., if the string
contains non-identifier characters or would be case-folded). Embedded quotes
are properly doubled.
»
B.t.w, the value of "quote_ident()" rests on the distinction between a name
(what you provide with the function's actual argument) and an identifier (what
it returns). Some of you flatly reject (borrowing a phrase from Tom) the
distinction between these two terms of art. Oh well…
Try this:
create table dog$(n int); -- OK
create table $dog(n int); -- Bad
create table "$dog"(n int); -- OK
These outcomes are consistent with the rules that say when a proposed name
needs to be double-quoted to form its identifier in a SQL statement (or
PL/pgSQL source text).
So it's correct for this to return FALSE:
select '$dog' = quote_ident('$dog');
But it's incorrect w.r.t. "quotes are added only if necessary" for this to
return FALSE:
select 'dog$' = quote_ident('dog$');
"format()" shows the same error when you use the %I placeholder. I suppose that
"format()" and "quote_ident()" share the same underlying implementation.
select format('What happens with %I?', 'dog'); -- double quotes are not added
select format('What happens with %I?', 'dog$'); -- double quotes are added
