# Select with prompt sends blank to validates\_inclusion\_of?

**URL:** <https://www.ruby-forum.com/t/select-with-prompt-sends-blank-to-validates-inclusion-of/141396>\
**Category:** Rails\
**Created:** [June 25, 2008, 12:51pm UTC](https://www.ruby-forum.com/t/select-with-prompt-sends-blank-to-validates-inclusion-of/141396 "2008-06-25T12:51:17Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![hozer](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/5fc32e/32.png) [@hozer](https://www.ruby-forum.com/u/hozer)\
**Post date:** [June 25, 2008, 12:51pm UTC](https://www.ruby-forum.com/t/select-with-prompt-sends-blank-to-validates-inclusion-of/141396/1 "2008-06-25T12:51:17Z")

</div>

Hi All, I hope somebody can help.

I’m a bit confused about what would be the best practice for handling an  
enumerated model value with FormOptionsHelper.select +  
validates\_inclusion\_of. Specifically the nil handling.

Some background…

I have an enumerated set of values, which I’m validating using  
validates\_inclusion\_of. Gender (like in the Rails docs example) is  
something I have to tackle, but additionally nil is allowed - as there  
will be no gender information supplied at initial user registration:

validates\_inclusion\_of :gender, :in =\> %w( m f ), :allow\_nil =\> true

An edit profile form will allow gender to be chosen later. In this form,  
I’m using FormOptionsHelper’s select with :prompt, so that a user is  
prompted to make a choice if they have not done so yet. i.e. the model’s  
underlying gender value is nil. For situations where a gender has  
already been chosen, just the male / female choice is preferred (hence  
:prompt seemed a better match to requirements than :include\_blank which  
always provides a blank option).

The problem is that the gender field “prompt” value gets passed through  
for the gender parameter as blank ("") when no selection is made, but  
I’d prefer the value to be left at nil. Rails implementation  
considerations aside, if I had a nil value in the model before, and made  
no change, I’d expect to retain a nil value after.

So, is :allow\_nil of more limited use in validations, for the case where  
no input field ever updates the value?

Is it better to opt for :allow\_blank =\> true on any validation where  
values are supplied via web form? (I take it that :allow\_blank accepts  
“” and nil as the blank? method does)

Does anybody go to the effort of translating blank parameter values back  
to nils after form submission - and is it worth it?

Any help would be gratefully received. I can’t actually believe I never  
stumbled over this one before! I guess enumerated values just feel a bit  
different from regular text values, as nil & blank never bothered me for  
those. I’m using Rails 2.1 & Ruby 1.8.7

Thanks!

---

<div class="post-metadata">

**Author:** ![hozer](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/h/5fc32e/32.png) [@hozer](https://www.ruby-forum.com/u/hozer)\
**Post date:** [June 27, 2008, 8:48am UTC](https://www.ruby-forum.com/t/select-with-prompt-sends-blank-to-validates-inclusion-of/141396/2 "2008-06-27T08:48:06Z")

</div>

Jim,

the choice about when to allow null in the database and how to handle  
nil in Rails is something I have been pondering of late. Sometimes  
there is a clear reason for doing it one way or another, but in  
general, it seems unnclear to me if there is any advantage either  
way.

It led me to wondering so I did a bit of googling and came up with  
this current discussion which discusses the issue of converting blank  
fields to nil - you may be interested in it. (I am bookmarking it for  
further thought).

[http://opensoul.org/2007/2/7/validations-on-empty-not-nil-attributes](http://opensoul.org/2007/2/7/validations-on-empty-not-nil-attributes)

sorry though - I still have not drawn any conclusions - I have been  
tending to allow nil on my database columns of late, but I think I  
have just convinced myself that that may not be the best approach.

I would be very interested to know if the core team have had any  
philosophical considerations about use of nil versus blank, and if so  
are there some Rails assumptions that are made, or has anything been  
documented. (perhaps it has and I just haven’t found it - I have the  
feeling of deja vu, but so much water goes under the bridge)

Tonypm