If you have the setup, give it a try.
On Jun 10, 6:59 pm, Thadeus Burgess wrote:
> Idea:
>
> Theoretically speaking of course if I were to create a virtual
> machine loaded with an apache instance and web2py on mod_wsgi. Using
> the various methods of *faking* an autonumber field on different
Idea:
Theoretically speaking of course if I were to create a virtual
machine loaded with an apache instance and web2py on mod_wsgi. Using
the various methods of *faking* an autonumber field on different
virtual machines, for the sake of scientific research.
Using the apache AB testing upon a
Understood. I will test this ASAP.
--
Thadeus
On Wed, Jun 9, 2010 at 11:05 AM, mdipierro wrote:
> This is not going to stay, it is just for you to look at.
> Consider defining the field as integer and create a trigger using SQL
> to autofill this field.
>
> On Jun 9, 10:52 am, mdipierro wro
This is not going to stay, it is just for you to look at.
Consider defining the field as integer and create a trigger using SQL
to autofill this field.
On Jun 9, 10:52 am, mdipierro wrote:
> well, I am posting in trunk a modifiled sql.py that
>
> allows Field('name','autoincrement') and generates
well, I am posting in trunk a modifiled sql.py that
allows Field('name','autoincrement') and generates the following code
for postgresql (only postgresql). Give it a try.
Massimo
On Jun 9, 10:38 am, Thadeus Burgess wrote:
> Postgres
>
> CREATE TABLE foo (
> id integer PRIMARY KEY SERIAL,
> bar
Postgres
CREATE TABLE foo (
id integer PRIMARY KEY SERIAL,
bar varchar,
did integer DEFAULT SERIAL);
Or...
CREATE SEQUENCE seq_foo_did START 2;
CREATE TABLE foo (
id integer PRIMARY KEY SERIAL,
bar varchar,
did integer DEFAULT nextval('seq_foo_did'));
SQLite however does not support multi
On a second thought of this matterthis is nonsense.
If a table has two autoincrement fields ("id" and "other") they always
contain the same values or values that differ for a constant offset
(If the started with different values). That is why some database
backends probably do not even support
I am not sure about the postgresql solution. Tell me how you do it in
SQL and i tell you how to do in web2pyese.
On Jun 9, 9:01 am, Thadeus Burgess wrote:
> Great. What about sqlite?
>
> --
> Thadeus
>
> On Wed, Jun 9, 2010 at 8:41 AM, mdipierro wrote:
> > In postgresql you get it native:
>
> >
I am testing this (both sqlite and postgresql). How do you do it in
raw sql?
This does not work in sqlite sql
CREATE TABLE a(
id INTEGER PRIMARY KEY AUTOINCREMENT,
d CHAR(512),
b INTEGER AUTOINCREMENT
);
It does not seem to like two AUTOINCREMENT fields (which is what I
thought).
O
Great. What about sqlite?
--
Thadeus
On Wed, Jun 9, 2010 at 8:41 AM, mdipierro wrote:
> In postgresql you get it native:
>
> Field('yourtfield',SQLCustomType('integer','SERIAL PRIMARY
> KEY',encoder=(lambda x: int(x)),decoder=(lambda x:x)))
>
> On Jun 9, 5:28 am, Thadeus Burgess wrote:
>> T
In postgresql you get it native:
Field('yourtfield',SQLCustomType('integer','SERIAL PRIMARY
KEY',encoder=(lambda x: int(x)),decoder=(lambda x:x)))
On Jun 9, 5:28 am, Thadeus Burgess wrote:
> That is the thing, its *almost* the same, but its not a true postgres
> sequence. Postgres already has ma
That is the thing, its *almost* the same, but its not a true postgres
sequence. Postgres already has many years of development making sure
their auto number works, why can't I just use that instead of trying
to hack around the limitations of a system?
I don't have a choice. I *must* have native su
If it were possible to do a SQL insert without the dummy filed this
almost the same as creating a sequence. web2py can create a table
without any field but the "id", but I do not do not how to do an
insert without any field value.
On Jun 8, 8:12 pm, Thadeus Burgess wrote:
> This *might* work. You
This *might* work. You are right, it is still horrible... It might be
*effectively* accomplishing the same thing that sequences do on
PostgreSQL, however I still wouldn't use it in production as it feels
"hacky". I already have to re-design this table, so I might as well do
it 100% right.
I never
I know this horrible but it does solve some of the problems...
db.define_table('whopper_seq',Field('dummy'))
db.define_table('yourtable',...
Field("whopper_id", "integer",compute=lambda r:
db.whopper_seq.insert(dummy=None))
...)
On Jun 7, 8:29 pm, Thadeus Burgess wrote:
> I have a problem.
>
>
On Jun 7, 11:39 pm, Thadeus Burgess wrote:
> Legacy systems =) The database was already in place as this,
> unfortunately when I migrated from access to postgres I kept the same
> schema as I was under a time crunch to get *something* running And
> now it bites me in the butt.
>
> The thing
By the way, this is for PostgreSQL that I need this. I think that
autoincrement needs to be implemented for each database type natively.
The *other* guys have autoincrement already... curious as to why it is
left out of web2py.
As always, the sooner the better >.<
--
Thadeus
On Mon, Jun 7,
Legacy systems =) The database was already in place as this,
unfortunately when I migrated from access to postgres I kept the same
schema as I was under a time crunch to get *something* running And
now it bites me in the butt.
The thing is... I cannot do this
``db(db.table.id==3).update(db.tab
I'm not even going to ask how you got in this situation :) Could you?:
0) Create a patch to Field for an autoincrement field type
1) Create a new table (whopper_temp) on your database with an id field
that is NOT set to auto-increment (yet)
2) Copy all records from the table in question to whopper
19 matches
Mail list logo
