Tom Lane wrote:
> View? Hmm, we know that views involving GROUP BY or aggregates
> don't work very well. We hope to fix those problems in 7.2, but
> right now there's probably not much that can be done about it.
> I'd recommend avoiding views that use those features for now.
Yes, I've just dis
Paul McGarry <[EMAIL PROTECTED]> writes:
>> Yup. ExecEvalAggref is supposed to fetch the result of a (previously
>> computed) aggregate function. It's unhappy because it's not finding
>> anything to return. This is a bug --- can we see a complete example
>> that causes it?
> The view that I am
I used the command below to alter a table and it took with no problem, but the
REFERENCES appears to have been ignored since I can put invalid numbers into
last_post_id. Are there plans to add the ability to specify column
constrainsts with ALTER TABLE ADD COLUMN in the future? For now, I suppo
Paul McGarry wrote:
> If I do a 'SELECT * from view_name' and get 8 rows of data, when I
> do a 'SELECT count(*) from view_name' I get 8 rows of the
> number 1 returned.
I've attached a script which will generate a bunch of tables and
a view then do a select * and select count(*) on that view.
Hi Tom,
Tom Lane wrote:
> > ERROR: ExecEvalAggref: no aggregates in this expression context
> > Does it mean that there aren't any aggregate in the expression
> > context when there should be.
> Yup. ExecEvalAggref is supposed to fetch the result of a (previously
> computed) aggregate function.
Paul McGarry <[EMAIL PROTECTED]> writes:
> Can someone tell me what this error message means:
> ERROR: ExecEvalAggref: no aggregates in this expression context
> Does it mean that there aren't any aggregate in the expression
> context when there should be.
Yup. ExecEvalAggref is supposed to fet
Hi,
I'll be glad if you describe more details about the problem. Is it a kind of
statistical analysis or what?
Omid Omoomi
>From: "Kermani, Bahram" <[EMAIL PROTECTED]>
>To: "'[EMAIL PROTECTED]'" <[EMAIL PROTECTED]>
>Subject: [SQL] Median
>Date: Fri, 30 Jun 2000 17:37:06 -0700
>
>Hello,
>
>I am t
just for the heck of it I decided to run the following in case it helped
rvbs=# explain select count(*) from PI_Keywords where keyword like
'peripherals & access%';
NOTICE: QUERY PLAN:
Aggregate (cost=35.98..35.98 rows=1 width=4)
-> Index Scan using pi_keywords_idx on pi_keywords (cost=0.0
> What i ment was, that I have information about workers, and I have to make
> report on them. Report should contain status about workers with the
> restriction , that the data is about last day of the month. I use perl as
> a scripting language, and how the hell should I know how many days were i
I've notice on certain queries, I was waiting a long time for a return so I
set out to troubleshoot something here's what I ran into...
When I do a explain on a select looking for 'peripherals & access' it uses
sequential scan but any other keyword uses index scan.
I've ran the vacuum analyze on P
> I want to get a list of data sources with the corresponding data,
> if the data exists, or with null, if the data doesn't. But anyway
> I need the data sources (all of them).
You want an outer join. Postgres doesn't have that yet, but you can
mimic it.
> I have also tried:
> select source_nam
Hello,
Can someone tell me what this error message means:
==
ERROR: ExecEvalAggref: no aggregates in this expression context
==
Does it mean that there aren't any aggregate in the expression
context when there should be. Does it mean there are and there
shouldn't be?
For that ma
On Wed, 5 Jul 2000, Gerhard Dieringer wrote:
> Jean-Marc Libs wrote:
> >...
> >I have also tried:
> >select source_name,data_value from source,data where data_source_id=source_id union
>select source_name,source_id,NULL from source,data
>
> >This is a bit better, in the sense that I get back al
Hello,
I
am trying to do a Median or Trimmed-mean operation in postgreSQL. I was
wondering if anybody knew how to do it. I appreciate it if you reply to my email
address.
Thanks,
Bahram
Kermani
[EMAIL PROTECTED]
Antti Linno wrote:
> Ok, if I want to get non-aggregat data in groups, I use order by. This
> group by seemed so logical though, but the fruit that u can't have, is
> usually the most sweet.
>
> New question, how to get the last day of month(order data by last day of
> month). And to prevent chai
Not quite sure what you mean, but how about this:
select date_part('day', 'Jul 01 00:00:00 2000 PDT'::datetime-1);
You must specify Jul for June, i.e. always the following month.
You could get around that too, but since I don't know why you
need to do what you asked, I'll leave it at this.
Tr
I thought, I finished that group by theme.
What i ment was, that I have information about workers, and I have to make
report on them. Report should contain status about workers with the
restriction , that the data is about last day of the month. I use perl as
a scripting language, and how the he
Saltsgaver, Scott wrote:
> Which procedural language is best for writing PostgreSQL functions:
> 'PL/pgSQL', 'PL/perl', or 'PL/Tcl'?
>
> I know that I can't do table creates in 'PL/pgSQL'?
>
> Is there an obvious choice as far speed and size?
PL/Tcl has the advantage that it can do the table
On Wed, 5 Jul 2000, Antti Linno wrote:
> New question, how to get the last day of month(order data by last day of
> month).
I'm sorry but I don't understand how you could order something by a value
(last day of month).
you can only order things by a field (e.g. a date field) or fields, in
incr
Ok, if I want to get non-aggregat data in groups, I use order by. This
group by seemed so logical though, but the fruit that u can't have, is
usually the most sweet.
New question, how to get the last day of month(order data by last day of
month). And to prevent chain letter from misunderstanding
No, you don't you could say:
SELECT workgroup, id
FROM job
GROUP BY workgroups;
On Wed, 5 Jul 2000, Antti Linno wrote:
> Ok, if I want to use group by, then I have to put every select field
> afterwards to group by. But I do want to group by one field.
> Hence
> select workgroup,id from job gro
Hi. I just came accross the big trouble.
" ERROR: Procedures cannot take more than 8 arguments"
Really?!? Does anybody know what to do to make it take 9 and more argments.
grishick
[EMAIL PROTECTED]
E.g.
select workgroup from job group by workgroup;
will that do?
Troy
>
> Ok, if I want to use group by, then I have to put every select field
> afterwards to group by. But I do want to group by one field.
> Hence
> select workgroup,id from job group by workgroup,id;
> would create groups o
You can use group by without having an aggregate operator but if you use
an aggregate you have to put the group by column int the select.
By the way what is after the AND? I hope it is not another workgroup
condition because it doesn't have much sense, I think. If you are
filtering different type
Antti Linno <[EMAIL PROTECTED]> wrote:
> I have table job, and I want to select data from it, grouping by
> workgroups. When I use
> select id,name from job where workgroup='top leaders' AND ... group by
> workgroup;
> When I want to group, I get
> ERROR: Illegal use of aggregates or non-gr
Ok, if I want to use group by, then I have to put every select field
afterwards to group by. But I do want to group by one field.
Hence
select workgroup,id from job group by workgroup,id;
would create groups of 1?
But I want to group by workgroups.
Sorry if my explanation is fuzzy.
A.
On Wed, 5 Jul 2000, Antti Linno wrote:
> I have table job, and I want to select data from it, grouping by
> workgroups. When I use
> select id,name from job where workgroup='top leaders' AND ... group by
> workgroup;
this query is probably incorrect: because of the "where workgroup=" the
"gro
I have table job, and I want to select data from it, grouping by
workgroups. When I use
select id,name from job where workgroup='top leaders' AND ... group by
workgroup;
When I want to group, I get
ERROR: Illegal use of aggregates or non-group column in target list
I'm confused, why does quer
Jean-Marc Libs wrote:
>...
>I have also tried:
>select source_name,data_value from source,data where data_source_id=source_id union
>select source_name,source_id,NULL from source,data
>This is a bit better, in the sense that I get back all I need, but there
>are too many lines: when there is da
Hi all,
I am trying to do something unusual (for me) and I can't quite
find the relevant part in the documentation (no mention to joins
in the insert section, or the index).
I want to get a list of data sources with the corresponding data,
if the data exists, or with null, if the data doesn't. B
30 matches
Mail list logo
