Re: PL/pgSQL: Incomplete item Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
Cc: Quan Zongliang <quanzongliang(at)yeah(dot)net>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: PL/pgSQL: Incomplete item Allow handling of %TYPE arrays, e.g. tab.col%TYPE[]
Date: 2024-01-04 21:02:07
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com> writes:
> Now, I think so this simple patch is ready for committers

I pushed this with some editorialization -- mostly, rewriting the
documentation and comments. I found that the existing docs for %TYPE
were not great. There are two separate use-cases, one for referencing
a table column and one for referencing a previously-declared variable,
and the docs were about as clear as mud about explaining that.

I also looked into the problem Pavel mentioned that it doesn't work
for RECORD. If you just write "record[]" you get an error message
that at least indicates it's an unsupported case:

regression=# do $$declare r record[]; begin end$$;
ERROR: variable "r" has pseudo-type record[]
CONTEXT: compilation of PL/pgSQL function "inline_code_block" near line 1

Maybe we could improve on that, but it would be a lot of work and
I'm not terribly excited about it. However, %TYPE fails entirely
for both "record" and named composite types, and the reason turns
out to be just that plpgsql_parse_wordtype fails to handle the
PLPGSQL_NSTYPE_REC case. So that's easily fixed.

I also wonder what the heck the last half of plpgsql_parse_wordtype
is for at all. It looks for a named type, which means you can do

regression=# do $$declare x float8%type; begin end$$;

but that's just stupid. You could leave off the %TYPE and get
the same result. Moreover, it is inconsistent because
plpgsql_parse_cwordtype has no equivalent behavior:

regression=# do $$declare x pg_catalog.float8%type; begin end$$;
ERROR: syntax error at or near "%"
LINE 1: do $$declare x pg_catalog.float8%type; begin end$$;
CONTEXT: invalid type name "pg_catalog.float8%type"

It's also undocumented and untested (the code coverage report
shows this part is never reached). So I propose we remove it.

That leads me to the attached proposed follow-on patch.

Another thing we could think about, but I've not done it here,
is to make plpgsql_parse_wordtype and friends throw error
instead of just returning NULL when they don't find the name.
Right now, if NULL is returned, we end up passing the whole
string to parse_datatype, leading to unhelpful errors like
the one shown above. We could do better than that I think,
perhaps like "argument of %TYPE is not a known variable".

regards, tom lane

Attachment Content-Type Size
improve-handling-of-composites-in-%TYPE.patch text/x-diff 4.6 KB

In response to


Browse pgsql-hackers by date

  From Date Subject
Next Message Jim Nasby 2024-01-04 22:08:28 Re: SET ROLE x NO RESET
Previous Message Andres Freund 2024-01-04 20:23:09 Re: Confine vacuum skip logic to lazy_scan_skip