Hi,
I have a table with column that must use sequence generated number
(for example in Postgres), like this obj_id:
table_Manager = Table( 'Manager', meta,
Column( 'obj_id', Integer, Sequence('obj_id_seq'), ),
Column( 'duties', type= String, ),
Column( 'name', type=
I am building a SQL query one step at a time in different parts of my
application. I am using MySQL 4.1
Currently, the query being automatically built is:
(SELECT doc_index.doc_id, doc_index.doc_type, doc_index.ref_id,
doc_index.create_date, doc_index.modified_date FROM doc_index) JOIN
snack
stmt = doc_type_table.select()
join_table = get_join_table('widgets')
stmt = stmt.join(join_table,
SA.and_(doc_type_table.c.ref_id==join_table.primary_key[0],
doc_type_table.c.doc_type=='widgets'))
join_table2 = get_join_table('spam')
stmt = stmt.join(join_table2,
Hi,
Well, I just happened to say model in my original message. I could have
said Tool to check whether a bunch of SA tables matches the database
tables.
That is what model_update.py does (at least, it tries to!) - so is it
something you'd consider including as an SA plugin?
Paul
On 2/1/07,
hallo.
AFAI understand from the source, in the graph used for topology
sorting in session.flush(), the nodes are representing the objects'
mappers. (and not the tables as in metadata.create_all() ).
Also, mappers which inherit from others, are replaced by their
bases/roots.
Edges in the
i am asking this, because so far i am considering both
table/foregn-key graph and mapper/relation graph same, thus applying
the mincut for the table/foregnkey (which sets fkey.use_alter=True)
also to the other (which sets respective relation.post_update=True).
And in 99% of cases, this works.
a join() by itself shouldnt be executed. you need to call select()
off of it (or use it inside the from_obj param of another select()).
notice its not the parenthesis that are the problem, its the lack of
an enclosing SELECT ... FROM.
On Feb 2, 5:04 am, jlowery [EMAIL PROTECTED] wrote:
I am
a circular relationship between two mappers means that a dependency
sort revealed that the mapper-level dependency graph contains cycles.
that portion of the sort is then thrown away and replaced with a
dependency sort where the nodes are individual object instances.
so sort #1 is mappers as
A mincut algorithm finds the minimal number of edges to cut in a
cycled graph so it becomes without cycles.
http://en.wikipedia.org/wiki/Max-flow_min-cut_theorem
i.e. applying such algorithm over the graph of table dependencies
(foregnkey), one gets some minimal number of foreign keys to make
As I understand the confusion here, Ive added a check for supports
execution in rev 2289. so that when you try to execute the join,
it will raise an error. originally i wasnt sure what would be
executable and what would not but im hoping its pretty well nailed
down at this point. so far its:
On Feb 2, 2007, at 2:49 AM, Andrew Stromnov wrote:
cities = cities.select(limit=1).execute().fetchall() works well with
this configuration, but leads to error in nregion =
places.select(places.c.name == iv2).execute()
if nregion = places.select(places.c.name == iv2).execute() used
first,
On 2/2/07, Michael Bayer [EMAIL PROTECTED] wrote:
are you saying that executing a particular statement is changing some
state on the local connection ?
Not exactly what I had in mind. Execution of first appeared statement
assigns (underlying ?) charset for all following queries for all
I have site in production right now that is using tsearch2. What I
did to accommodate the results with SA was to simply not map the
tsearch2 column to the SA object. I have a view that creates the
tsvector objects based on the source table. (I actually created a
materialized view and indexed
thats a bug, i fixed it in r2291.
if you dont want to use the trunk for now, you can probably define
the column as:
Column('obj_id', integer, default=func.obj_id_seq.nextval())
On Feb 2, 2007, at 4:05 AM, che wrote:
Hi,
I have a table with column that must use sequence generated number
On Feb 2, 2007, at 10:33 AM, svilen wrote:
A mincut algorithm finds the minimal number of edges to cut in a
cycled graph so it becomes without cycles.
http://en.wikipedia.org/wiki/Max-flow_min-cut_theorem
i.e. applying such algorithm over the graph of table dependencies
(foregnkey), one
right, except that when you say statement1.execute() and then
statement2.execute(), its very likely that those two statements are
using distinct connections from the connection pool. also, SA is not
issuing any kind of charset anything on a connection so it would not
be within SA's
Hi,
Michael Bayer написа:
if you dont want to use the trunk for now, you can probably define
the column as:
Column('obj_id', integer, default=func.obj_id_seq.nextval())
this answers another question of mine :) that i planned to ask
...but it generates (on Postgres) this:
SELECT
i think you are looking for a feedback arc set, which describes
exactly the problem that applying post_update to a mapping
solves:
http://en.wikipedia.org/wiki/Feedback_arc_set
and we actually generate a set like this in the _find_cycles()
method, except it doesnt find just one edge of
oh sure, if not on the wiki directly of course it can linked/
downloaded/whatever from the site...id like to have as many recipes
and things as possible (which is really what the wiki is for).if
it were with the dist, id have it in a util or contrib folder off
the root.
On Feb 2,
oh right...try func.nextval('obj_id_seq')
On Feb 2, 2007, at 12:46 PM, che wrote:
Hi,
Michael Bayer написа:
if you dont want to use the trunk for now, you can probably define
the column as:
Column('obj_id', integer, default=func.obj_id_seq.nextval())
this answers another question of
On Feb 2, 2007, at 2:14 PM, svilen wrote:
of course, topological sort is a sort, but in order of it to work, the
graph should not have cycles. Which are cut by use_alter's and
post_updates, on foreign keys and relations.
well, in our case all edges have cost of 1, except
Sorry for the rather newbie question, but I can't find anything about
this anywhere else.
How do people writing db applications with SA do unit testing on their
code? Are there any good db unit testing frameworks that work well
with SA? If I can't find anything I will probably just roll my own
Allen schrieb:
How do people writing db applications with SA do unit testing on their
code? Are there any good db unit testing frameworks that work well
with SA? If I can't find anything I will probably just roll my own
but I would like to keep from reinventing the wheel if I can help it.
23 matches
Mail list logo