Question

This is a problem I have run into before and I have yet to find an elegant solution, so I thought I would ask for SO’s help.

I am iterating through an array and printing off some information from that array and having trouble figuring out how to print my start and end <div> tags. Below is an example table and the desired output along with a basic implementation of my current algorithm. I am hoping that somebody can point me to a better algorithm for printing the data. I’m doing it in PHP, but a generic algorithm would be wonderful.

Thanks very much!

Table: (spacing added for clarity)

Choice ID     Choice Body     Question ID     Question Body
---------------------------------------------------------------------
1             Yes, very much  1               Do you like sandwiches?
2             Somewhat        1               Do you like sandwiches?
3             Not at all      1               Do you like sandwiches?
4             I hate them     1               Do you like sandwiches?

5             Sure, why not   2               Do you like apples?
6             Yesh, I guess   2               Do you like apples?
7             What are those  2               Do you like apples?

8             Yes, very much  3               Do you like chips?
9             Not at all      3               Do you like chips?

Desired Output:

<div class='question' id='1'>
  <p>Do you like sandwiches?</p>

  <div class='choices'>
    <span class='choice'>Yes, very much</span>
    <span class='choice'>Somewhat</span>
    <span class='choice'>Not at all</span>
    <span class='choice'>I hate them</span>
  </div>
</div>

<div class='question' id='2'>
  <p>Do you like apples?</p>

  <div class='choices'>
    <span class='choice'>Sure, why not</span>
    <span class='choice'>Yeah, I guess</span>
    <span class='choice'>What are those</span>
  </div>
</div>

<div class='question' id='3'>
  <p>Do you like chips?</p>

  <div class='choices'>
    <span class='choice'>Yes, very much</span>
    <span class='choice'>Not at all</span>
  </div>
</div>

Basic Algorithm I’m Currently Using:

$last_id = null;
while ($choice = pg_fetch_array($choices)) {
    if ($last_id != $choice['id']) {
        if ($last_id != null) {
          echo "</div>";
        }

        echo "<div id='$choice[id]'>";
    }

    // Print choice info

    $last_id = $choice['id'];
}

if ($last_id != null) {
    echo "</div>";
}

Note: The reason I’m using this way is for optimization purposes. This requires only one database query, and there are going to be a lot of results. I don’t want to have to do a query for each question to get it’s choices. I know how to do that, and it is easier, but not very efficient or fast.

Edit 1: Fixed code, the algorithm now works, but still isn’t pretty. For the commenter: pg_fetch_array() is a PostgreSQL function that basically creates an associative array. Very similar to an object. Allows you to just ask for the $choice['id'] or $choice['body'].

Answer 1

Grouping items by similar values can hardly be called an algorithm. It’s more of a coding pattern if anything else.

A good way to code is to separate the mechanism from the intent. In this case the mechanism is how to keep track of the key values to find the grouping boundaries and the intent is to output HTML for each sequential group.

Python for instance has a library function called groupby to do exactly this. So in Python the code would look something like this (ignoring the fact that one would use a templating library for this):

from itertools import groupby

def question_from_row(row):
    return dict(id=row['question_id'], body=row['question_body'])

for question, choices in groupby(questions, key=question_from_row):
    print('<div class="question" id="%s">' % question['id'])
    print('  <p>%s</p>\n' % question['body'])
    print('  <div class="choices">')
    for choice in choices:
        print('<span class="choice">%s</span>' % choice['choice_body'])
    print('  </div>')
    print('</div>')

PHP to my knowledge doesn’t have anything like that built in, but a naive implementation is pretty easy:

function array_groupby($input, $keyfunc) {
    $output = array();
    $last_key = null;
    $current = null;
    foreach ($input as $item) {
        $item_key = $keyfunc($item);
        if ($last_key === null || $item_key != $last_key) {
            if ($current !== null) {
                $output[] = $current;
            }
            $last_key = $item_key;
            $current = array();
        }
        $current[] = $item;
    }
    if ($current !== null) {
        $output[] = $current;
    }
    return $output;
}

This would be typical library code that you include in. The code that deals with the output then becomes rather trivial. Completely isolated from how the grouping is done. For instance you could change array_groupby to return an object that implements the Iterator interface and only lazily fetches from the input iterable.

$questions = array_groupby(pg_fetch_array($choices),
    function($choice) { return $choice['id']; });

foreach ($questions as $choices) {
    $question = $choices[0];
    echo '<div class="question" id="'.$question['id'].'">';
    echo '<p>'.$question['body'].'</p>';
    echo '<div class="choices">';
    foreach ($choices as $choice) {
        echo '<span class="choice">'.$choice['choice_body'].'</span>';
    }
    echo '</div>';
    echo '</div>';
}

This example is using the PHP 5.3 closures feature. On older versions specifying the function to extract the grouping key would be slightly uglier, perhaps calling for an object oriented approach.