Implicit cast of literal in SQL statements

From: Scott Royston <scroyston71(at)yahoo(dot)com>
To: pgsql-bugs(at)postgresql(dot)org
Subject: Implicit cast of literal in SQL statements
Date: 2001-12-27 20:37:14
Message-ID: 830E7B29-FB09-11D5-92C0-000393577144@yahoo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs pgsql-hackers

I've seen a few postings in multiple newsgroups saying that in 7.1.x and
up, literals in SQL statements are implicitly cast to strings.

For example in:
select distinct 'hello' from mytable;
the 'hello' is implicitly assumed to be 'hello'::text

However, in both 7.1.3, and a fresh build of 7.2b4 from cvs, (with all
regressions passing) I get:

mytest=# select distinct 'hello' from mytable;
ERROR: Unable to identify an ordering operator '<' for type 'unknown'
Use an explicit ordering operator or modify the query

an explicit 'hello'::text works fine.

I've spent a day looking through the code and can't really find any
obvious #define's or compile time flags that would be causing this
problem.
It looks like
Const *
make_const(Value *value)
{
...
case T_String:
val = DirectFunctionCall1(textin,
CStringGetDatum(strVal(value)));

typeid = UNKNOWNOID; /* will be coerced
later */
typelen = -1; /* variable len */
typebyval = false;
break;
...
}

does the damage, and it never gets 'coerced later', at least not before
transformDistinctClause(...) gets called, which is where the failure
happens (a few levels down).

does this really work for everybody else? Can someone point me to a
compile flag I may be missing, or the code that actually does the
implicit cast?

thanks

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Tom Lane 2001-12-27 21:22:56 Re: Implicit cast of literal in SQL statements
Previous Message Tom Lane 2001-12-27 14:49:50 Re: Serializable access giving wrong error messages?

Browse pgsql-hackers by date

  From Date Subject
Next Message Pavlo Baron 2001-12-27 20:37:40 Re: patch: INSERT INTO t VALUES (a, b, ..., DEFAULT, ...)
Previous Message Pavlo Baron 2001-12-27 16:19:51 Re: patch: INSERT INTO t VALUES (a, b, ..., DEFAULT, ...)