New Access Method

Alan Gatt wrote:

Hello,

I am trying to create a new index which is multidimensional (based on
R-Tree) which can be used for data warehosuing. I have read the Developers'
manual about adding new indexes, and now I want to start coding the new
index.

My first step is to create the skeleton files, but I cannot compile
Postgres
now because of errors. This is what I have done:
- Create a header file in /src/includes/
- Create a folder in /backend/access/ for the new index.
- Created a new C file in the above folder
- Copied a Makefile from the other access methods, and changed the
references to point to my C file
- Edited the /backed/access Makefile to include my new folder

I am using Eclipse on WindowsXP. Any pointers about what have I done wrong
or anything I missed?

Well you haven't given us any errors or what c compiler you using...

Joshua D. Drake

Thanks,
Alan

--

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

On Sat, Mar 03, 2007 at 07:38:15PM +0100, Alan Gatt wrote:

Hello,

I am trying to create a new index which is multidimensional (based on
R-Tree) which can be used for data warehosuing. I have read the Developers'
manual about adding new indexes, and now I want to start coding the new
index.

As stated you havn't given any details as to the actual problem.

However, you don't need to change the source tree at all to make a new
index on a type. You usually make the helper functions and then create
the operator class and you're done...

Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/

Alan Gatt escribi�:

Hello,

I am trying to create a new index which is multidimensional (based on
R-Tree) which can be used for data warehosuing. I have read the Developers'
manual about adding new indexes, and now I want to start coding the new
index.

Have you considered coding it using the GiST infrastructure? Should be
much easier, and automatically give crash safety, concurrency, and a
shorter development time.

--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.

Ok, so I am using GCC 3.2 as a compiler, and the following is the error
message:

make[4]: Entering directory
`/c/dev/postgres/pgsql_tip/src/backend/access/mdim'
gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Winline
-Wdeclaration-after-statement -Wendif-labels -fno-strict-aliasing -g
-I../../../../src/include -I./src/include/port/win32 -DEXEC_BACKEND
"-I../../../../src/include/port/win32" -DBUILDING_DLL -c -o mdim.o mdim.c

make[4]: *** [mdim.o] Error 3
make[4]: Target `SUBSYS.o' not remade because of errors.
make[4]: Leaving directory
`/c/dev/postgres/pgsql_tip/src/backend/access/mdim'
make[3]: *** [mdim-recursive] Error 2

If I choose the GIST development, do you know of any documentation which can
help me?

Thanks,
Alan

Also, any pointers where I can find the operator classes and functions for
the GIST index?

Thanks

On Sun, Mar 04, 2007 at 09:32:58AM +0100, Alan Gatt wrote:

Also, any pointers where I can find the operator classes and functions for
the GIST index?

The documentation is a good start:

http://www.postgresql.org/docs/8.1/static/gist.html

There are also plenty of examples in the contrib portion of the source
code.

Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/

"Alan Gatt" <alangatt@gmail.com> writes:

make[4]: Entering directory
`/c/dev/postgres/pgsql_tip/src/backend/access/mdim'
gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Winline
-Wdeclaration-after-statement -Wendif-labels -fno-strict-aliasing -g
-I../../../../src/include -I./src/include/port/win32 -DEXEC_BACKEND
"-I../../../../src/include/port/win32" -DBUILDING_DLL -c -o mdim.o mdim.c

make[4]: *** [mdim.o] Error 3

Umm ... what happened to the actual compiler error message? I'd have
expected to see something in between those lines.

regards, tom lane