Thread: [PgFoundry] Unsigned Data Types
Hello all,
Here is the first pass at the unsigned data type I have been working on.
I am planning on adding these to the September 2008 commitfest wiki page.
The unsigned data type is not targeted for core, but for the uint PgFoundry project.
The uint.c.gz file is the main source file for the uint1, uint2, and uint4 data types.
The uing.sql.gz file contains the SQL statements to add the unsigned data type to the database.
The pg_atoui.c.gz file is based off the function in the PostgreSQL source code but works for unsigned data types instead of signed data types.
The Makefile is used to build the unsigned data type shared library on Linux.
The tests.tar.gz is my unit test suit that I worked on to make sure the unsigned integer types worked as expected.
The tests cover cases like:
* table creation with the unsigned integer types.
* comparision operations.
* INSERT statements (binary and text forms).
* COPY statements (binary and text forms).
* unique btree index support.
In addition to correctness issues, I would also appreciate feedback on best practices and portability concerns.
For example:
I doubt my Makefiles are very portable.
What is the proper solution to handle this? pgxs?
Thanks,
- Ryan
Here is the first pass at the unsigned data type I have been working on.
I am planning on adding these to the September 2008 commitfest wiki page.
The unsigned data type is not targeted for core, but for the uint PgFoundry project.
The uint.c.gz file is the main source file for the uint1, uint2, and uint4 data types.
The uing.sql.gz file contains the SQL statements to add the unsigned data type to the database.
The pg_atoui.c.gz file is based off the function in the PostgreSQL source code but works for unsigned data types instead of signed data types.
The Makefile is used to build the unsigned data type shared library on Linux.
The tests.tar.gz is my unit test suit that I worked on to make sure the unsigned integer types worked as expected.
The tests cover cases like:
* table creation with the unsigned integer types.
* comparision operations.
* INSERT statements (binary and text forms).
* COPY statements (binary and text forms).
* unique btree index support.
In addition to correctness issues, I would also appreciate feedback on best practices and portability concerns.
For example:
I doubt my Makefiles are very portable.
What is the proper solution to handle this? pgxs?
Thanks,
- Ryan
Attachment
On Aug 15, 2008, at 1:00 AM, Ryan Bradetich wrote: > Here is the first pass at the unsigned data type I have been > working on. > > I am planning on adding these to the September 2008 commitfest wiki > page. > The unsigned data type is not targeted for core, but for the uint > PgFoundry project. Is the intention for the types to go into pg_catalog? It'd be nice if you could specify what schema they should be installed in. An uninstall would also be good. Thanks for doing this, I've wished we had uint types in the past, and I'm sure I will again in the future! -- Decibel!, aka Jim C. Nasby, Database Architect decibel@decibel.org Give your computer some brain candy! www.distributed.net Team #1828
On Sat, Aug 16, 2008 at 10:53 AM, Decibel! <decibel@decibel.org> wrote:
The pg_catalog made since to me at first (especially for my application), but on reflection I believe you are right. I will remove the references to the pg_catalog schema and allow the user to add the unsigned data type to any schema. Good catch on the uninstall script. I should have written this as well. I will post an update to the wiki later tonight.
I am glad it is useful. I needed it for my current project, and I was hoping others could use it as well.
Thanks,
- Ryan
On Aug 15, 2008, at 1:00 AM, Ryan Bradetich wrote:Is the intention for the types to go into pg_catalog? It'd be nice if you could specify what schema they should be installed in. An uninstall would also be good.Here is the first pass at the unsigned data type I have been working on.
I am planning on adding these to the September 2008 commitfest wiki page.
The unsigned data type is not targeted for core, but for the uint PgFoundry project.
The pg_catalog made since to me at first (especially for my application), but on reflection I believe you are right. I will remove the references to the pg_catalog schema and allow the user to add the unsigned data type to any schema. Good catch on the uninstall script. I should have written this as well. I will post an update to the wiki later tonight.
Thanks for doing this, I've wished we had uint types in the past, and I'm sure I will again in the future!
Thanks,
- Ryan
I can say that we have had several times to use bigint instead because of the lack of uint type in postgres.
On Sun, Aug 17, 2008 at 9:03 PM, Ryan Bradetich <rbradetich@gmail.com> wrote:
On Sat, Aug 16, 2008 at 10:53 AM, Decibel! <decibel@decibel.org> wrote:I am glad it is useful. I needed it for my current project, and I was hoping others could use it as well.On Aug 15, 2008, at 1:00 AM, Ryan Bradetich wrote:Is the intention for the types to go into pg_catalog? It'd be nice if you could specify what schema they should be installed in. An uninstall would also be good.Here is the first pass at the unsigned data type I have been working on.
I am planning on adding these to the September 2008 commitfest wiki page.
The unsigned data type is not targeted for core, but for the uint PgFoundry project.
The pg_catalog made since to me at first (especially for my application), but on reflection I believe you are right. I will remove the references to the pg_catalog schema and allow the user to add the unsigned data type to any schema. Good catch on the uninstall script. I should have written this as well. I will post an update to the wiki later tonight.
Thanks for doing this, I've wished we had uint types in the past, and I'm sure I will again in the future!
Thanks,
- Ryan