Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:42382
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: error (pb1.pair.com: domain pooteeweet.org from 88.198.8.16 cause and error)
Cc: internals@lists.php.net
Message-ID: <03D6160A-376F-44A8-B82F-F31873CFA0BB@pooteeweet.org>
To: Mike Panchenko <m@mihasya.com>
In-Reply-To: <5faa7b820812260106p6f24bf0ct49ef574e5ed4e2c2@mail.gmail.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v930.3)
Date: Fri, 26 Dec 2008 16:56:16 +0100
References: <5faa7b820812260106p6f24bf0ct49ef574e5ed4e2c2@mail.gmail.com>
Subject: Re: [PHP-DEV] Proposal: array_flatten function
From: mls@pooteeweet.org (Lukas Kahwe Smith)


On 26.12.2008, at 10:06, Mike Panchenko wrote:

> Hey everybody, I'm new both to the list and to hacking the  
> internals, so
> I'll try to keep it short and humble.
>
> I've written an array_flatten function which just takes all elements  
> from a
> nested array and pushes them into a single indexed array.
>
> Couple of questions:
>
> 1. Most importantly, is this patch-worthy to everybody? I was a bit
> surprised that it didn't already exist (it wasn't difficult to  
> write, even
> for me :) ), and a search turns up quite a few posts both on the PHP  
> doc
> comments and elsewhere detailing how to do the same in usersland  
> PHP. There
> are some pretty clear usecases for this, such as doing statistics on a
> grouped dataset, as well as grabbing all values from a single-column  
> SQL
> query. It's a very simple, unobtrusive patch.
>
> 2. One of the PHP implementations "preserved keys." This seems
> counterintuitive to me, as there's a high probability of  
> overwriting, and I
> can't think of a usecase where one would want to flatten out an  
> array, but
> keep the keys. Nonetheless, it was implemented by one of the  
> snippets I saw.
> Should the function offer this option with a $preserveKeys = false  
> optional
> param? If so, how should I handle duplicate keys? Failure? Warning?  
> IMO,
> it's not worthwhile, but I wanted to get opinions.
>
> 3. (this is where the noob comes in) What is the protocol on creating
> patches? Should I just run the CVS patch command against a major  
> revision? I
> haven't used that before, but it seems straight forward enough. If  
> someone
> can point me to a writeup of the exact "correct" way to generate a  
> patch,
> I'd appreciate it.

In one of my recent projects I had the need for something like this. I  
essentially had a recursive data structure of organizations and  
frequently I had to get a list of all organization ID's that were  
above or underneath a given organization. So I had to write a deeply  
self join which would produce lets say around 3-4 columns that either  
contained an ID or were NULL.

I ended up with an implementation like the following. Actually looking  
at it now it seems way too complex but it works, so it goes:
     function flattenArray($array) {
         if (empty($array)) {
             return array();
         }
         $first = reset($array);
         if (empty($first)) {
             return array();
         }
         $keys = array_keys($first);
         $result = array();
         foreach ($array as $row) {
             foreach ($keys as $key) {
                 if (!empty($row[$key])) {
                     $result[] = $row[$key];
                 }
             }
         }
         $result = array_unique($result);
         return $result;
     }

So I guess what I am asking for is a flag to ignore NULL's. Kind of  
reminds me of that patch Igor proposed the other day. Then again, we  
have sooooo many array functions its not even funny. So maybe we  
really should look more towards Iterators to help us consolidate  
things. Not sure.

regards,
Lukas