But Paolo -- you forgot the third option...
FIX the problem in dal.py so that Validator classes work properly. It has
no deleterious effects on other validators (at least so far in my testing)
and it permits validators to format None values.
-- Joe B.
On Friday, October 26, 2012 9:56:57 AM UTC-7, Paolo Caruccio wrote:
>
> In my opinion, you have mainly two options to bypass None value check that
> formatter function in DAL does:
>
> option 1 : store None in database, delete formatter from validator, use
> represent in table field (I prefer this approach)
>
> class IS_ELAPSED_TIME(object):
> def __init__(self,error_message='Must be MM:SS.hh or MMSShh (with no
> punctuation)'):
> self.error_message=error_message
> def __call__(self,value):
> try:
> if value and value.upper() != 'NT':
> res = hms_to_int(value)
> else:
> res = None
> return (res,None)
> except:
> return (value,self.error_message)
>
>
> Field('est', 'string', requires=IS_ELAPSED_TIME(), represent=lambda value,
> row:int_to_hms(value) if value else "NT")
>
>
> option 2: store 'NT' in database in place of None
>
> class IS_ELAPSED_TIME(object):
> def __init__(self,error_message='Must be MM:SS.hh or MMSShh (with no
> punctuation)'):
> self.error_message=error_message
> def __call__(self,value):
> try:
> if value and value.upper() != 'NT':
> res = hms_to_int(value)
> else:
> res = 'NT'
> return (res,None)
> except:
> return (value,self.error_message)
> def formatter(self,value):
> if value != 'NT':
> rtn = int_to_hms(value)
> else:
> rtn = 'Elapsed Standard Time not available'
> return rtn
>
>
> Field('est', 'string', requires=IS_ELAPSED_TIME())
>
>
>
> Il giorno venerdì 26 ottobre 2012 04:13:55 UTC+2, Joe Barnhart ha scritto:
>>
>> I do manage the "None" value in my formatter, Paulo.
>>
>> But it never gets called because the built-in code of the Field class
>> tests the value for None and then exits before calling my formatter. So it
>> doesn't matter that my formatter can handle None -- it is never called.
>>
>> That's why I posted the change to the Field class in the first message of
>> this thread.
>>
>> -- Joe B.
>>
>> On Thursday, October 25, 2012 5:53:43 PM UTC-7, Paolo Caruccio wrote:
>>>
>>> Why don't you manage None values in int_to_hms function?
>>>
>>>
>>> Il giorno giovedì 25 ottobre 2012 03:20:29 UTC+2, Joe Barnhart ha
>>>>>> scritto:
>>>>>>>
>>>>>>> I have an application where I expect "None" items in my database and
>>>>>>> I want to format them to "NT". It is an app that uses time standards,
>>>>>>> and
>>>>>>> if there is no standard present I expect a "None" in the database which
>>>>>>> translates to a field of "No Time" or "NT".
>>>>>>>
>>>>>>> The problem is that the current implementation of formatter in the
>>>>>>> Field class tests the value for "None" and escapes before the formatter
>>>>>>> is
>>>>>>> called.
>>>>>>>
>>>>>>> I can see why this behavior might be expected in a lot of cases, but
>>>>>>> it seems extreme to deny the ability to format "None" into a more
>>>>>>> pleasing
>>>>>>> form for those applications that could benefit from it. Here is the
>>>>>>> offending part of formatter (located in gluon/dal.py):
>>>>>>>
>>>>>>> def formatter(self, value):
>>>>>>> requires = self.requires
>>>>>>> if value is None or not requires:
>>>>>>> return value
>>>>>>>
>>>>>>> If I change the above to:
>>>>>>>
>>>>>>> def formatter(self, value):
>>>>>>> requires = self.requires
>>>>>>> if not requires:
>>>>>>> return value
>>>>>>>
>>>>>>> I get my desired behavior, which is to pass "None" to my formatter
>>>>>>> which is implemented as part of a custom Validator object. I realize
>>>>>>> the
>>>>>>> code now has to go "further" for cases where the value is None, but is
>>>>>>> it
>>>>>>> really safe to assume nobody will ever want to "format" None into
>>>>>>> another
>>>>>>> string? Not in my case, at least!
>>>>>>>
>>>>>>> Joe B.
>>>>>>>
>>>>>>>
>>>>>>>
--
