Обсуждение: column definition for setof record functions (like dblink and crosstabs)
Hi,
Functions that return SETOF RECORD need to be called with a column definition. That is a bit limiting.
For example, you might not know how many columns your crosstab will have in advance, you might not know the structure of a table that you access with dblink in advance.
Also, it's always a more complicated query to write, which is not very user friendly.
I searched a bit for the reason, and i found: "This allows the system to understand what * should expand to, and what proname in the WHERE clause refers to, in advance of trying to execute the function." in the dblink page.
From this sentence i can derive that there are some internal mechanisms that need to have the the column definition before the data.
I was just wondering if there were a way to generate this information automatically, so that the user need not supply it.
It would have a cost of course, but in some cases, it might certainly be worth it.
Cheers,
WBL
--
"Quality comes from focus and clarity of purpose" -- Mark Shuttleworth
Functions that return SETOF RECORD need to be called with a column definition. That is a bit limiting.
For example, you might not know how many columns your crosstab will have in advance, you might not know the structure of a table that you access with dblink in advance.
Also, it's always a more complicated query to write, which is not very user friendly.
I searched a bit for the reason, and i found: "This allows the system to understand what * should expand to, and what proname in the WHERE clause refers to, in advance of trying to execute the function." in the dblink page.
From this sentence i can derive that there are some internal mechanisms that need to have the the column definition before the data.
I was just wondering if there were a way to generate this information automatically, so that the user need not supply it.
It would have a cost of course, but in some cases, it might certainly be worth it.
Cheers,
WBL
--
"Quality comes from focus and clarity of purpose" -- Mark Shuttleworth
2012/6/6 Willy-Bas Loos <willybas@gmail.com>: > Hi, > > Functions that return SETOF RECORD need to be called with a column > definition. That is a bit limiting. > For example, you might not know how many columns your crosstab will have in > advance, you might not know the structure of a table that you access with > dblink in advance. > Also, it's always a more complicated query to write, which is not very user > friendly. > > I searched a bit for the reason, and i found: "This allows the system to > understand what * should expand to, and what proname in the WHERE clause > refers to, in advance of trying to execute the function." in the dblink > page. > From this sentence i can derive that there are some internal mechanisms that > need to have the the column definition before the data. > > I was just wondering if there were a way to generate this information > automatically, so that the user need not supply it. > It would have a cost of course, but in some cases, it might certainly be > worth it. > there are no way, how to do it. generating cross tables should be solved via cursors http://okbob.blogspot.cz/2008/08/using-cursors-for-generating-cross.html I believe so dblink will be deprecated in near future, because has mentioned limits - you can use http://wiki.postgresql.org/wiki/Foreign_data_wrappers > Cheers, > > WBL > -- > "Quality comes from focus and clarity of purpose" -- Mark Shuttleworth >