it? If Access' forms interface is so badly broken that few
people use it anyway, then I'd say that we should not break standards
compatibility just to support it.
regards, tom lane
---(end of broadcast)---
TIP 1: subscribe
ature that sends out COMMITs for you.
If you have that turned on, *and* you are explicitly committing, then
that's probably the cause of the extra commits. Get rid of one or
the other.
regards, tom lane
plications for.
At one time Bruce had made some patches to emit informative notice
messages about implicit FROM entries, but that got turned off again
for reasons that I forget...
regards, tom lane
failed to notice that it was giving you undefined
results.
regards, tom lane
l into your
database? It's not installed by default ...
regards, tom lane
is* covered in the FAQ, see
http://www.postgresql.org/docs/faq-english.html#4.22
regards, tom lane
he process quit, or just hang up
waiting for something? Is there anything reported in the postmaster
log? (You might try firing up the psql run with environment variable
PGOPTIONS = "-d2" to cause individual queries to be logged.)
regards, tom lane
Bruce Momjian [EMAIL PROTECTED] writes:
I can confirm this is still a bug.
Yeah, it's a parser problem. I punted it in Lockhart's direction, but
he's been out of town...
regards, tom lane
oh, btw this select refused to use an SQL natural join among the 3
relations
Tulassay Zsolt [EMAIL PROTECTED] writes:
FATAL 1: cannot create init file pg_internal.init
Hmm, disk full? Wrong ownership or permissions on database
directory (data/base/tek)?
Looks like that error report needs to include the kernel
error message...
regards, tom
regards, tom lane
the
explicit conversion:
regression=# create function int2(int8) returns int2 as
regression-# 'begin return int2(int4($1)); end;' language 'plpgsql';
CREATE
regression=# insert into t2 (select b from t1 where a = 1);
INSERT 314649 1
regards, tom lane
*) palloc(16)) != NULL)
strcpy(result, password-password);
return (result);
}
That doesn't return "text", so you can't tell the system it does.
Type text is a varlena type, ie, length word followed by data.
regards, tom lane
two dates is probably
what you need to do.
I don't think we have a two-input MAX function like that, but it'd be
easy enough to fake it with a CASE expression. Something like
... ORDER BY (CASE WHEN a b THEN a ELSE b END) DESC;
regards, tom lane
et
whether they are kernel I/O requests or logical I/O requests, ie which
side of the buffer cache they are counted on.
regards, tom lane
expression requires its
very own parentheses, so you need two sets:
sum((select ))
But the other way is probably better...
regards, tom lane
-nothing proposition.
regards, tom lane
the backend,
PQendcopy() failed. Explanation from backend: 'FATAL 1: Memory
exhausted in AllocSetAlloc()
'.
The query was: 'COPY "medusa" WITH OIDS TO stdout;
Hmm. What is the full definition of that table? (pg_dump -s -t medusa
would do.)
regards, tom lane
Ryan Bradetich [EMAIL PROTECTED] writes:
Tom Lane wrote:
Ryan Bradetich [EMAIL PROTECTED] writes:
-- dumping out the contents of Table 'medusa'
FATAL 1: Memory exhausted in AllocSetAlloc()
PQendcopy: resetting connection
SQL query to dump the contents of Table 'medusa' did not execute
if there is a restriction clause matching the index. In your
example you have only a join WHERE clause.
Plain btree indexes on city.country_id and country.country_id
might work better --- at least they'd offer the option of
a merge join without doing explicit sort.
regards, tom lane
e evidence that another plan
type would be quicker for this problem?
regards, tom lane
the oldstyle cases beyond 8
are likely to be dead code anyhow ... not much point in cluttering the
backend with more cases.)
I think I need to change the config.h comment that goes with
MAX_FUNC_ARGS.
regards, tom lane
Thomas Swan [EMAIL PROTECTED] writes:
I'm at a loss on how identify or get the names of children of a parent class.
Join pg_class against pg_inherits.
regards, tom lane
of the actual
runtimes.
regards, tom lane
it
happen?
regards, tom lane
, tom lane
plpgsql's fault, but that of the function-call
interface. The problem appears no matter what programming language
you write the function in.
regards, tom lane
, or 7.1 if
we get lucky.
regards, tom lane
, tom lane
be done about it.
I'd recommend avoiding views that use those features for now.
regards, tom lane
need the backslash trick.)
regards, tom lane
need to match a literal backslash
in a LIKE pattern, how many backslashes do you have to write in your
query?
regards, tom lane
g at a lot more data to be fetched by
the scan, not just 10 records, so the indexscan becomes less attractive.
Might be interesting to compare the estimated and actual runtimes
between this query and what you get with "set enable_seqscan to off;"
regards, tom lane
s then
you *are* giving him Postgres superuser rights. He can do anything
he damn well pleases as user postgres, he's just got to write some
code to do it.
Once you've absorbed that, giving him the dbadmin password should
not look unreasonable.
regards, tom lane
to go about it. I'd think about making a table containing a
pattern column and a browser-name column and doing the view as a
join. Might be a little tricky to ensure you get only one match,
however ...
regards, tom lane
s
no error message of that spelling in 7.0 pg_dump,
but there is in 6.5 ...
Check your PATH, or invoke pg_dump using a full path
instead of letting it default.
regards, tom lane
tate, "NEW" is planned new row state
DELETE: "OLD" holds row to be deleted
regards, tom lane
d function marked "iscachable"
(for details see the archives for the last time this question came up,
a few months back). But the lowest-tech solution may be to calculate
the cutoff time on the application side, so you can just send it as a
constant to begin with.
regards, tom lane
of a view is that it's not an
exact image of the underlying data, so how is a machine going to
figure out what you want an update on the view to do?
regards, tom lane
...
regards, tom lane
fied
value.
I'd suggest folding the log-entry-making into your trigger, actually.
If you have a trigger anyway then the insert into t1_log is only one
more line in that trigger procedure...
regards, tom lane
:-(. Thus, for example,
select * from foo where random() 0.5
doesn't work as desired. I plan to fix this for 7.1.
Offhand I don't see a use for a nextval-like function in WHERE,
so you're probably safe with both 6.5 and 7.0.
regards, tom lane
INCT might help if so.)
regards, tom lane
apps that do it all the time. I suspect
maybe the plpgsql parser gets confused by SQL-style comments? Your
example seems to work fine if I leave out the comment.
regards, tom lane
before. Perhaps
7.0.* is broken in this respect? Would think people would have noticed,
though.
regards, tom lane
it to be or not. This has been gone over *many*
times before on this mail list, so I didn't think I needed to repeat it.
This will be fixed in 7.1 (is already fixed in current sources).
regards, tom lane
limitation so far has been bogus
assumptions in the function-call API, not querytrees.
regards, tom lane
an SQL function
instead of plpgsql, if that helps any...
regards, tom lane
's caching of query plans (and
specifically the table OIDs stored in those plans) or is there another
issue here?
We do need to think about invalidating cached query plans when updates
happen...
regards, tom lane
SIX 1003.2 regexps
(see src/backend/regex/re_format.7). I believe there is an item on the
TODO list about upgrading the regexp parser to something more modern
... feel free to hop on that project if it's bugging you ...
regards, tom lane
of how to define a column default that gives the
time of insertion. You need to write this as
NEW.ChangedAt := now();
to prevent the system from reducing timestamp('now') to a constant
when the function is first executed.
regards, tom lane
on a regular basis that
doesn't seem like it'd be the issue. Anyway, keep an eye on things with
ipcs and we'll see what's going on...
regards, tom lane
.
regards, tom lane
keep track of multiple changes
intra-transaction). It might not be defined quite the way you want,
but if you're not picky about what a "version number" is, it'll do.
regards, tom lane
. Hopefully by 7.2 there
will be a better way to do the view.
regards, tom lane
d)"
and I get this result:
ERROR: rewrite: comparision of 2 aggregate columns not supported
I think this is fixed in 7.0.
regards, tom lane
inserting?
I've recently been swatting some performance problems in the btree
index code for the case of large numbers of equal keys.
regards, tom lane
column is
mistakenly counting the hidden IdSort column that's needed to sort by.
I don't know of any good workaround in 7.0, short of patching the
erroneous code. Have you thought about using a view, rather than a
function returning set?
regards, tom lane
ULL except for requiring a boolean
input, AFAICT).
regards, tom lane
want that.
regards, tom lane
to punctuation.
That's pretty bizarre (not to say difficult to believe). What LOCALE
setting are you running the postmaster in?
regards, tom lane
:
regression=# select 968518585 :: int4 :: abstime :: timestamp;
?column?
2000-09-09 12:56:25-04
(1 row)
(there's probably a cleaner way to do this, but that works ...)
regards, tom lane
und JOIN expressions...
regards, tom lane
Meszaros Attila [EMAIL PROTECTED] writes:
Can I test this feature in the current snapshot?
Sure. But see my message to pghackers on Tuesday for notes about what's
not working yet in the JOIN support.
regards, tom lane
Keith Wong [EMAIL PROTECTED] writes:
Anybody know how to compile pgaccess from postgres source files?
I think it's driven by configure --with-tcl.
regards, tom lane
"Michael Richards" [EMAIL PROTECTED] writes:
It appears that postgres 7.0 does not support repeatable read for
transaction isolation. Is this planned? If so, when?
?? Maybe you need to do SET TRANSACTION ISOLATION LEVEL?
regards, tom lane
table
declarations, so I can try the example without a lot of guessing?
regards, tom lane
Louis-David Mitterrand [EMAIL PROTECTED] writes:
On Wed, Sep 20, 2000 at 10:43:59AM -0400, Tom Lane wrote:
Louis-David Mitterrand [EMAIL PROTECTED] writes:
How should I interpret that error?
ERROR: replace_vars_with_subplan_refs: variable not in subplan target list
Looks like a bug to me
like a bug to me ... Thomas, do you agree?
Curiously, it appears that the interval input converter will accept this
with or without the 'days' keyword included.
regards, tom lane
instead of 'Not a view' when there is no view by the
given name, but I do not agree with trying to suppress errors due to
metadata problems.
regards, tom lane
Meszaros Attila [EMAIL PROTECTED] writes:
How far is the above subject from beeing implemented?
I'm looking at it right now ... no promises yet, though.
regards, tom lane
maybe egcs is misoptimizing around the inline
assembly code of tas() ?
regards, tom lane
. It'd just be another way to trip yourself up when
moving across DBMSes.
regards, tom lane
);
?
regards, tom lane
OIDs).
2. counter overflow problems hit you only per-table, not
per-installation.
regards, tom lane
over just as many tables as you need to have unique IDs
across. That way you don't have a problem at dump/reload time,
and you don't exhaust your ID space any faster than you must.
regards, tom lane
active
table, per-table sequence values look like a better bet.
BTW, there *is* talk of providing an 8-byte-OID option, but I'm not
holding my breath for it.
regards, tom lane
article.id
and articles_groups.groupid, also. You seem to have indexes on all
the wrong columns of articles :-( ... each of those indexes costs you
on updates, but will it ever be useful in a query?
regards, tom lane
In general any type
coercion can be provided this way --- a function of a single argument,
having the same name as its return type, represents a coercion path.
regards, tom lane
f term that can mean different things to different people...
regards, tom lane
don't need superuser help to do this, you just have to do
GRANT ALL ON table TO yourself
as the table owner. But it's stupid to have to do that when it's
supposed to be the default condition. Fixed for 7.1.
regards, tom lane
...
regards, tom lane
ly got any permissions granted to himself;
ie, changing permissions is not a grantable/revokable right,
it's just checked on the basis of who you are.
regards, tom lane
a snapshot installation is probably a waste of time.)
regards, tom lane
or might not have a sensible
default behavior for types it doesn't recognize.
regards, tom lane
.
regards, tom lane
...
regards, tom lane
SQL around here, not Perl, so we have to follow
the SQL spec's definition of substr().
regards, tom lane
, but they're all
there.
For an UPDATE, you could check to see whether old.col = new.col.
This would miss the case where an UPDATE command is explicitly setting
a column to the same value it already had; dunno if you care or not.
regards, tom lane
e works OK in current sources,
but until 7.1 comes out you'll need to write something like
where login.login = lower('foo')::char;
Or change the login field to type text...
regards, tom lane
rk a lot
better when there are large numbers of rows involved. It might actually
happen for 7.2...
regards, tom lane
pitting up on a stray \r.
Dunno what your admin did to make the problem appear where you
hadn't had it before...
regards, tom lane
cast
There are no comparison operators for array types ...
regards, tom lane
cast
There are no comparison operators for array types ...
regards, tom lane
ransaction :-(. So it chokes when you access
more than 1000 LOs in the same transaction.
Dunno about your other issue, but clearly your application is
failing to report whatever error message was returned when the
transaction was aborted...
regards, tom lane
Pierre Habraken [EMAIL PROTECTED] writes:
It looks like if subqueries in from clause are not supported by
PostgreSQL. Am I right ? If yes, are there any plans to provide this
feature soon ?
Already there in current sources for 7.1 ...
regards, tom lane
Roberto Mello [EMAIL PROTECTED] writes:
I was wondering if UNIONs in VIEWS will be supported too.
Already there for 7.1.
regards, tom lane
undesirable
hidden side effects.
Yes, and they won't be too hidden either: it won't work :-(
Current sources check for that sort of type mismatch, but 6.5 failed to
do so.
regards, tom lane
r System V
IPC semaphore configuration. Form more information
Most likely your kernel isn't set to allow shared memory blocks as
large as 8M.
regards, tom lane
column is added to the children too, as
it should be, but in an unexpected column position, which causes trouble
for pg_dump (a dump and reload will do the wrong thing). Perhaps what
you heard is a garbled report of that issue.
regards, tom lane
type 501"
AFAICT it works in 7.0.* as well. Are you sure this was 7.0.2, and
not 6.5.something?
regards, tom lane
1 - 100 of 2222 matches
Mail list logo
