hum.. interesting... just got what you meant by dal vs sql.... now
that I think of it, Massimo did mention that dal was a 2nd rev of sql
and that it was experimental. but, that said. I have been using dal,
not sql. I did have to tweak a few things to make it work, but in
general, works great for me.
I have not specified neither migrate nor fake_migrate. I just know
that dal will make use of them when creating table
(t._create(migrate=migrate, fake_migrate=fake_migrate) Note: I think
there is an equivalent t.create (no under score but would have to
check how they differ). Note, about 1/2 the tables I create are
created on the fly, and the table_name will depend on how some
upstream variables get resolved (can change a lot)...
One mistake I kept making that kept causing exceptions (until my brain
was sufficiently caffeinated) was when creating some Objects (of
whatever type, again on the fly), I kept making the mistake of using a
var name for the table_name, previously used elsewhere where that name
had served to name a new reference to a class (sorry for the
ridiculous sentence (sounds better in french ;)) - I think I meant I
re-used var names).
myVar = "generated_class" <-- something generates this string
ref_to_class_generated = myVar <-- "a" is just the name
...
later instantiate new class
myVar = name_of_ref_to_class_previously_generated
myVar = initNewObj(data)
myVar.DoStuff()
bla bla bla
myVar = str(some_object) --> 'className_is_now_a_string'
so later if I did something like this:
tableName = myVar
db.define_table(tableName,
SQLField('name'),
SQLField('value'))
I got that bad table name exception, then it went through that
migrate,fake_migrate bit.
So I re-use the name but I tag it for the tableName with a '_name', so
they don"t get confused.
I.e.
myVar_name = str(some_object) -->
'className_is_still_a_string_but_this_worked'
had no issues since.
Mart :)
On Oct 20, 11:12 am, Thadeus Burgess <[email protected]> wrote:
> In my experience the dal.py does not work stand alone, however sql.py does.
>
> Table migrations have always worked for me when using standalone.
>
> --
> Thadeus
>
> On Wed, Oct 20, 2010 at 9:58 AM, Bruno Rocha <[email protected]> wrote:
> > did you specified both migrate and fake_migrate ?
>
> > 2010/10/20 mart <[email protected]>
>
> > forgot to mention something a well...
>
> >> I think the issue I had was related to yours with the migration,
> >> because creating a table, without specifying migrate= produces the
> >> following exception while defining a table. That migration data as
> >> well as the parameters I passed in both get validated by
> >> t._create(migrate=migrate, fake_migrate=fake_migrate). This is why I
> >> think migrating or creating tables with no migration... both are
> >> subject to the same rules, risking the same exceptions.
>
> >> db.define_table(tableName,
> >> SQLField('blueModuleStr'),
> >> SQLField('blueModuleObj','blob'),
> >> SQLField('blueModuleImports'))
>
> >> objMakeDB.instModule(folder)
> >> File "/Users/mart/Documents/Aptana Studio Workspace/blueLite/src/
> >> blueLite/pyModules/createModuleTable.py", line 34, in instModule
>
> >> SQLField('blueModuleImports'))
> >> File "/Users/mart/Documents/Aptana Studio Workspace/blueLite/src/
> >> blueLite/pyUtils/gluon/dal.py", line 1399, in define_table
>
> >> t._create(migrate=migrate, fake_migrate=fake_migrate)
> >> File "/Users/mart/Documents/Aptana Studio Workspace/blueLite/src/
> >> blueLite/pyUtils/gluon/dal.py", line 1869, in _create
>
> >> Mart :)
>
> >> On Oct 19, 7:11 pm, mart <[email protected]> wrote:
> >> > I have recently introduced the web2py DAL to some back-end stuff so
> >> > that it would play well with the front end (web2py). Although I did
> >> > trim it down and the amount of files in the gluon folder (I bootstrap
> >> > for each start of each software build, so size matters) and got rid of
> >> > some unresolved imports caused by the triming (i don't need web access
> >> > here, just the dal). So, are you taking about where (path) the .db and
> >> > tables get created? if this is the case, then I found 2 things:
>
> >> > 1) the db and tables don't seem to follow the same rule in that the db
> >> > can get created just about anywhere, where the tables seem to get
> >> > created relative to where *db.define_table(tableName,...)* is called
> >> > (seems to be the default). so depending on where you are in the
> >> > structure... also, I notice I had to be xtra sensitive with error
> >> > handling in that, if a previous step failed to lets say do an update
> >> > or an insert and if I didn't handle that well at THAT moment, then the
> >> > next time that field was referenced (which caused an exception), it
> >> > create the entire set of default tables I setup and would do so where
> >> > ever the module doing the EXECUTE would be. Which lead to look at
> >> > dal.py
>
> >> > 2)so, her, the code can be changed to modify that behavior, and I kept
> >> > good focus while following the flow of the script, but it is
> >> > relatively large file, and I didn't take notes as I was reading. But
> >> > it should be doable. the trick is to isolate the code directly related
> >> > to 1) the adapter of the of the db your are using and the table/and
> >> > migration related actions (that's where we see most of the references
> >> > to the folder housing the tables). I haven't tried yet, and i don"t
> >> > know if doing this would offend Massimo, so I held back and stuck with
> >> > being relative to the folders where I generate tables.
>
> >> > BTW - i believe this is the code causing your exception, so one of
> >> > your params is not in line with what's expected ("if not in key") or
> >> > its type is wrong (just guessing though).
>
> >> > for key in args:
> >> > if key not in [
> >> > 'migrate',
> >> > 'primarykey',
> >> > 'fake_migrate',
> >> > 'format',
> >> > 'trigger_name',
> >> > 'sequence_name']:
> >> > raise SyntaxError, 'invalid table "%s" attribute: %s'
> >> > % (tablename, key)
>
> >> > hope it helps.
>
> >> > Mart :)
>
> >> > On Oct 19, 3:37 pm, Bruno Rocha <[email protected]> wrote:
>
> >> > > Somebody knows a trick?
>
> >> > > 2010/10/19 Bruno Rocha <[email protected]>
>
> >> > > > I forgot to mention that I tried:
>
> >> > > > DAL(....,folder=...) pointing folder="" to the directory where
> >> .table
> >> > > > files are, but does not works.
>
> >> > > > 2010/10/19 Bruno Rocha <[email protected]>
>
> >> > > > I know DAL was not made for that, but I'm using the DAL in a desktop
> >> > > >> application with PyGTK, and it is working very well :-)
>
> >> > > >> It is a simple application that monitors the presence of employees
> >> in a
> >> > > >> company and reads small CSV files from a time clock,
> >> > > >> people has cards that open the gates/doors of the company factory,
> >> I use a
> >> > > >> stream to read the track from serial port of time clock,
> >> > > >> then, I take the information serialized as CSV, I parse and write
> >> it into
> >> > > >> SQLite db, after that , the Janitor uses a PyGTK app to access that
> >> > > >> information.
>
> >> > > >> already been running for about 6 months, So far everything is
> >> working
> >> > > >> fine, but I can not run the automatic migrations.
>
> >> > > >> Does anyone know a way to make migration work automatically with
> >> DAL Stand
> >> > > >> Alone?
>
> >> > > >> I'm importing sql.py I'm connecting with SQLite, setting tables,
> >> accessing
> >> > > >> and doing out any crud operation.
>
> >> > > >> The only thing missing is to make migration works.
>
> >> > > >> I already set migrate='Mytable.table' and I tried with migrate=True
>
> >> > > >> ----
> >> > > >> An example of what I have working in my
>
> >> > > >> "connect.py"
> >> > > >> >>> from gluon.sql import *
> >> > > >> >>> db = DAL('sqlite://timeclock1.db')
> >> > > >> >>> Track =
>
> >> db.define_table('track',Field('regnumber','integer'),Field('action','integer'),Field('timestamp','datetime'),migrate='track.table')
>
> >> > > >> "Form_workflow.py"
> >> > > >> >>> Track.insert(regnumber=123,action=2,timestamp='2010-10-19')
> >> > > >> 1
> >> > > >> >>> Track.insert(regnumber=124,action=2,timestamp='2010-10-19')
> >> > > >> 2
> >> > > >> >>> db.commit
>
> >> > > >> Until here, its ok.
>
> >> > > >> But now I am wanting to change the model, and including
> >> > > >> Field('department')
>
> >> > > >> "connect.py"
> >> > > >> >>> Track =
>
> >> db.define_table('track',Field('regnumber','integer'),Field('action','integer'),Field('timestamp','datetime'),
> >> > > >> *Field('department')*,migrate='track.table')
>
> >> > > >> Traceback (most recent call last):
> >> > > >> File "<stdin>", line 1, in <module>
> >> > > >> File "/bin/DAL/gluon/sql.py", line 1346, in define_table
> >> > > >> raise SyntaxError, 'invalid table name: %s' % tablename
> >> > > >> SyntaxError: invalid table name: track
>
> >> > > >> ----
>
> >> > > >> If this is not possible, I'll have to create new fields in SQLite
> >> and then
> >> > > >> update my model.
>
> >> > > > --
>
> >> > > >http://rochacbruno.com.br
>
> >> > > --
>
> >> > >http://rochacbruno.com.br
>
> > --
>
> >http://rochacbruno.com.br
>
>
