# Many-to-many and error handling

**URL:** <https://www.ruby-forum.com/t/many-to-many-and-error-handling/117303>\
**Category:** Rails\
**Created:** [October 15, 2007, 8:49pm UTC](https://www.ruby-forum.com/t/many-to-many-and-error-handling/117303 "2007-10-15T20:49:06Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![sp00nyg](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/edb3f5/32.png) [@sp00nyg](https://www.ruby-forum.com/u/sp00nyg)\
**Post date:** [October 15, 2007, 8:49pm UTC](https://www.ruby-forum.com/t/many-to-many-and-error-handling/117303/1 "2007-10-15T20:49:06Z")

</div>

I have several simple many-to-many relationships in my app. Some use  
a join model, and others don’t. Let’s start with one that doesn’t.

class Company \< ActiveRecord::Base  
has\_and\_belongs\_to\_many :industries

# Add an industry if it isn’t already there

def add\_industry(name)  
name.strip!  
unless self.industries.find\_by\_name(name)  
industry = Industry.find\_or\_initialize\_by\_name(name)  
self.industries \<\< industry  
end  
end  
end

I have three tables: companies, industries, and companies\_industries.

The problem I have is best practices to handle errors with habtm and  
Ruby. There are several things that could go wrong here. First,  
there should be an update lock on the initial find to ensure that two  
processes don’t try to add the same industry. Does :lock =\> true work  
properly here? Normally, I use that where I want to lock a single  
row, but that’s not what is happening here.

The next problem is how to get errors from self.industries \<\<  
industry. Both industries and company\_industries are inserted into  
with this statement. I usually try to use create! and save! in case  
there is some db error, but I don’t have that option here. Won’t this  
fail silently or by returning nil? If it returns nil, I do I know  
what failed?

What are the best practices for making this bulletproof? Note, I’m  
old-school enough want to know what is going on under the covers and  
protect my app from it.

Thanks!

---

<div class="post-metadata">

**Author:** ![sp00nyg](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/edb3f5/32.png) [@sp00nyg](https://www.ruby-forum.com/u/sp00nyg)\
**Post date:** [October 15, 2007, 10:21pm UTC](https://www.ruby-forum.com/t/many-to-many-and-error-handling/117303/2 "2007-10-15T22:21:04Z")

</div>

On 15 Oct 2007, at 19:48, gobigdave wrote:

> ```
> unless self.industries.find_by_name(name)
> 
> ```
> 
> there should be an update lock on the initial find to ensure that two  
> processes don’t try to add the same industry. Does :lock =\> true work  
> properly here? Normally, I use that where I want to lock a single  
> row, but that’s not what is happening here.

A unique index on companies\_industries can be used to enforce the  
uniqueness. :lock =\> true locks an individual record, but that’s not  
what you want here. Depending on your database there may be other  
things you can do with regard to locking.

> The next problem is how to get errors from self.industries \<\<  
> industry. Both industries and company\_industries are inserted into  
> with this statement. I usually try to use create! and save! in case  
> there is some db error, but I don’t have that option here. Won’t this  
> fail silently or by returning nil? If it returns nil, I do I know  
> what failed?

Looking at the source, the return value of the call to \<\< will be  
false if somethng funny happened.

Fred

---

<div class="post-metadata">

**Author:** ![sp00nyg](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/s/edb3f5/32.png) [@sp00nyg](https://www.ruby-forum.com/u/sp00nyg)\
**Post date:** [October 15, 2007, 11:28pm UTC](https://www.ruby-forum.com/t/many-to-many-and-error-handling/117303/3 "2007-10-15T23:28:27Z")

</div>

I have the unique index, but it looks like self.industries \<\< industry  
doesn’t give me a way to know why it returned false. That’s too bad.  
It would be nice to raise an exception with more information.

Is it standard practice in rails to ignore these errors? Previous web  
apps I did a lot with stored procs so I had more control and info.

On Oct 15, 4:19 pm, Frederick C. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)