Re: shortest way to recover from a QuerySet "object not found"

> Just create your own Manager and override the default (named 'objects') in > your models. Have 'get' behave any way you like. > > http://docs.djangoproject.com/en/1.1/topics/db/managers/ > > Shawn Ding! http://docs.djangoproject.com/en/1.1/topics/db/managers/#adding-extra-manager-methods luv

Re: shortest way to recover from a QuerySet "object not found"

> The only condescension I've seen in this thread is from you. And, to > be fair, if I wanted to be condescending I'd have simply pointed you > at Tony Hoare's explanation of null values That's why I said NullObject, in the first post. -- You received this message because you are subscribed to t

Re: Adding a new convenience filter 'contains_any'

Peter Herndon wrote: > Won't the "in" filter do exactly what you need? > > http://docs.djangoproject.com/en/1.1/ref/models/querysets/#in Why do you need even __in? Given... Entry.objects.filter(id__in=[1, 3, 4]) ...couldn't id=[] overload if the target is

Re: Adding a new convenience filter 'contains_any'

;list's) in fields, the =[] event is free to be "overloaded" by its type. So id=[] would resolve to id__in=[], id=Scalar to id__exact=Scalar, and the explicit versions are available if you suspect that overloading is fragile. (I also disagree with leaving out the spaces around =, but obvi

Re: Adding a new convenience filter 'contains_any'

On Mar 5, 8:41 am, Peter Herndon wrote: > On Mar 4, 2010, at 7:37 PM, Phlip wrote: > > (I also disagree with leaving out the spaces around =, but obviously > > obeying a team style guide supersedes improving it...) > > ... the goal of clarity/legibility uh... -- You

Re: Adding a new convenience filter 'contains_any'

> >>> (I also disagree with leaving out the spaces around =, but obviously > >>> obeying a team style guide supersedes improving it...) > > >> ... the goal of clarity/legibility > > > uh... > > ;)  I meant *having* a style guide was consistent with said goal. right - following an aesthetic style g

Re: Adding a new convenience filter 'contains_any'

> In this case, it's not just a team style guide - it's PEP8, which clearly > says: And if the PEP8 told you to ... just jump off a cliff... would you? -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django

Re: Adding a new convenience filter 'contains_any'

> > And if the PEP8 told you to ... just jump off a cliff... would you? > > Sounds like you might benefit from actually reading it: This thread arc is about me lamenting the positive value of style guides, even those with wrong line-items. Nobody here has said to get creative with any style guide

SNIPPET: assert_model_changes(record, 'field', 41, 42, lambda: money_line())

eval() a string instead of serve an attribute? self.assert_model_changes( blog, 'post.count()', 41, 42, lambda: blog.write_one_post('yack yack yack') ) The code shows how to .reload() in Django, because (unless I'm wrong), Django models don't ha

def reload(model): model.__dict__ = model.__class__.objects.get(pk=model.pk).__dict__ ; return model

ort (or fail to prevent) assigning a .__dict__ like that? -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-us...@googlegroups.com. To unsubscribe fr

Re: def reload(model): model.__dict__ = model.__class__.objects.get(pk=model.pk).__dict__ ; return model

iously - reload(instance) does nothing to the instance we already have, and its internals are still dirty... -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send emai

Re: PyCharm IDE - Anyone using it yet

ecades do we have to wait until editors support testing as aggressively as debugging? -- Phlip http://penbird.deviantart.com/art/Uncle-Wiggily-Wants-You-156969773 -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this

Re: Negate querysets

> Just create another queryset that excludes everything in your first > queryset: > > negated_queryset = User.objects.exclude(id__in=queryset.values("id")) QuerySets are already so easy to plug-n-play... Ain't there a way to do it without whacking the database twice?

Re: Negate querysets

"SELECT * WHERE id not in (SELECT id WHERE ...)" might just snocker the database worse than two SELECTs. You never know; that's why I'd like to see an EXPLAIN on it! -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Googl

Re: Negate querysets

> I want to be able to answer your question, forever. Is there a way, > from a developer test, to query "what SQL statement does this QuerySet > generate"? Oh, duh, it's QuerySet.query, as a string. Brand X makes that one inconceivably hard, due to poor factoring... >

assert_performant(query_set)

at, I have used this technique, copiously, on other projects, and whenever that assertion failed it indeed saved us from accidentally deploying a slow SQL statement to our server! -- Phlip http://zeekland.zeroplayer.com/Pigleg_Too/1 -- You received this message because you are subscribed to t

Re: assert_performant(query_set)

orrect, we will pin down their behavior if we parse back the intermediate value - the actual SELECT statement, and then interrogate it. Oh, yeah, and such a test is also perform...ing. Well. Because it doesn't build records or read them. I will see when I have time to code that up... -- Phlip

new_records = assert_latest(query_set, lambda)

#x27;s only one - we sample the lambda source, to make the error message as clear as possible; this works best if the lambda's on its own line - every assert needs a deny; I will write deny_latest as soon as a need appears. -- Phlip http://zeekland.zeroplayer.com/Pigleg_Too/

automatically import lots of modules at python manage.py shell time?

. How can we fix the first problem? How can this (otherwise useful) shell use a minimal or invisible import rubric, to grab, say, a bunch of models so they are all ready & available for use? -- Phlip http://penbird.deviantart.com/ -- You received this message because you are subscribed to

Re: automatically import lots of modules at python manage.py shell time?

local_dict[model_class.__name__] = model_class del local_dict, appcache, os ''') os.system('ipython -i -nobanner .ipython') -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users"

Re: Testing email with the testserver

;s why the documentation assumed you started there. -- Phlip -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-us...@googlegroups.com. To unsubscribe from this group, send email to django-us

Re: assert_performant(query_set)

> qs.query.get_compiler(qs.db).as_sql() That gives this error message with the usual zero Google hits for that error message: AttributeError: 'BaseQuery' object has no attribute 'get_compiler' -- You received this message because you are subscribed to the Google Groups "Django users" group.

Re: assert_performant(query_set)

>   AttributeError: 'BaseQuery' object has no attribute 'get_compiler' qs._as_sql() returns a tuple of a SELECT statement, and a (), presumably with wildcards; I will start there. -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this

Re: assert_performant(query_set)

>  * Similarly, connection.cursor() will only work in the single > database case. connections[qs.db].cursor() Aaaand that's not working either. (No connections, no qs.db.) I must take a break until someone unravels all this, because I don't know enough about the architecture inside QuerySet... -

Re: How to share settings file in multiprogrammer project

e one copied up to the server. All sub-settings files should start with from settings import * to pull in the common settings, and should override them if they need to. One important override is test_settings should use sqlite3 :memory: database. -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You

Re: How to share settings file in multiprogrammer project

> The test_settings and dev_settings should point to a PASSWORDLESS > database with localhost-only permissions. The production_settings file > should not be committed. It's the one copied up to the server. Come to think of it, what I'm screaming about here is three separate things: - developer d

Re: I am having issues with django test

>     def setUp(self): >         user = User.objects.create_user(USER, '@nowhere.com', PASS) That's what fixtures are for! -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-us...@googlegroups.com. T

Re: assert_performant(query_set)

do with Django - is ... cursor.execute('EXPLAIN ' + query) ...seems to scramble a sqlite3 database. If I take the EXPLAIN out I don't get unrelated test failures (tho the new assertion naturally does not work!). I will make it work for MySQL! -- Phlip -- You received this message becaus

assert_compliant_sql

tput The source appears below my sig (where I am aware archivers might trim it). Warning: Per my other thread, calling EXPLAIN on an a SQLite3 connection seems to scramble it. Not sure why, but I don't care, because one should not use SQLite3 in production. -- Phlip http://c2.com

Re: How do I properly unit test a Django session?

E_NAME] = store.session_key session = self.client.session session.update(dictionary) session.save() # and now remember to re-login! -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users

manage.py loaddata silently fails

Djangoists: (Under version 1.1.1) loaddata (and the fixtures= line in tests) silently fails if the data cannot load. How do I access the error message? -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django

Re: manage.py loaddata silently fails

warning. The fixtures= line also has this bug. -- Phlip -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-us...@googlegroups.com. To unsubscribe from this group, send email to

what lines in the templates are culpable?

Djangoists: When code below a template throws an error, we get an insanely detailed stack trace of all the lines AROUND the template. How do I tell what lines INSIDE the templates caused the error? -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are

Re: what lines in the templates are culpable?

> Uusually the first line in the traceback tells you pretty explicitly > where the error is. Without the traceback, it's hard to say where your > problem lies. One of the templates is "basket.html", and "basket.html" does not appear in the transcript. All the lines are only django's internal rend

Re: what lines in the templates are culpable?

> If you posted some specifics of what you are actually seeing, someone might > be able to help. I mean, actually copy-paste the traceback into your post > (or put it someplace like dpaste.com and point to it). I am not asking "oo help I can't get my template working". That part's done - by clampi

how to get the test runner to only rebuild & populate the test_database if it changed?

ere an easier way, or has someone already done this? (1.2, BTW - the obese database prevents an upgrade to Django 1.3!) -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this grou

Re: how to get the test runner to only rebuild & populate the test_database if it changed?

" database. It does not even change the file time on the .db file when the tests run. Could someone better at Django internals than me package this system up and contrib it? -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Grou

cascading create

to optimize the database calls. If not, that's okay to. Google doesn't say Django can do this. My specific problem is I have a dozen trees with a couple thousand leaves each, and simply writing them all causes a bottleneck. --   Phlip   http://c2.com/cgi/wiki?ZeekLand -- You rec

Re: cascading create

Bump? String together the SQL myself? On Jan 31, 3:40 pm, Phlip wrote: > Djangoists: > > Given a model Tree with many Leaves, I want to write this: > >   t = Tree(data=42) >   t.leaves.add(leafy_data=43) >   t.leaves.add(leafy_data=44) >   t.save() > > I want the s

How to pivot a table?

n re-pack the array with a for-loop. Do QuerySet aggregations and annotations offer some way to push that query into the database? -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users" group.

Re: Unit Testing: temporarily altering django settings

o trivially pass it, without real user input. To test the captcha, mock it to pass or fail. -- Phlip http://c2.com/cgi/wiki?ZeekLand -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this group, send email to django-users@

order_by clause to preserve order in list? - redux

s.google.com/group/django-users/browse_thread/thread/f23b2bdf24ff7a26/110d3e5c9e69ca63 I will, of course, start with extra(), and see where that takes me! -- Phlip http://zeekland.zeroplayer.com/ -- You received this message because you are subscribed to the Google Groups "Django users"

Re: Unit Testing: temporarily altering django settings

On Feb 21, 12:47 pm, Cody Django wrote: > Thanks -- I didn't know about mock objects, and this is good to know. > But this doesn't feel like this is the solution I'm looking for.  It's > a large project, and your proposal would require extensive patching. Does your project access the CAPTCHA set

Re: Select x random rows from DB

On Feb 20, 2:19 pm, galago wrote: > What is the best way, to select X random rows from DB? I know that > method: .all().order_by('?')[:X] is not good idea. > What methods do you use? order_by('RAND()') That might use the same seed each time. To create, for example, a rotating home page with

Re: order_by clause to preserve order in list? - redux

eniences" as the .order_by() builder. It insisted on parsing my string instead of simply appending it. > -- >   Phlip >  http://zeekland.zeroplayer.com/ -- You received this message because you are subscribed to the Google Groups "Django users" group. To post to this grou

Field forced to not-null

the null=True. Any idea why this is happening? I have deleted the complete Django project and started from scratch only to find the same scenario; when I click save, I get the following message. Any advice will be welcomed. Regards, Phlip -- You received this message because you are subs

Re: Field forced to not-null

Thank you Andréas and Jason That solved my problem Regards, Phlip On Monday, August 6, 2018 at 3:02:01 PM UTC+2, Andréas Kühne wrote: > > Hi Philip, > > The field is required because you haven't allowed it to be blank - and > django admin requires fields to have the fol

Recursive one-to-many foreign key on the same table

== I get the following error message: no such column: komadm_apps_decisionindex.IndexPredecessor_idno such column: Any help will be welcome. Regards, Phlip -- You received this message because you are subscribed to the Google Groups "Django users" group.

Re: Recursive one-to-many foreign key on the same table

Thank you Marius; the video is worth the watch and thought provoking Phlip On Thursday, August 9, 2018 at 2:42:13 PM UTC+2, Phlip Pretorius wrote: > > I want to build a one-to-many recursive foreign key on a table in Django. > > The concept is something like this: > > T

<    1   2