Making an index for a search by PHP

Question

How can you search only unique words with PHP such that I can learn the basics in making the search?

I have had a few problems in making a multidimensional array for questions.

My first unsuccessful attempt is the following.

#1

$result = pg_query_params ( $dbconn, 
    "SELECT question_id, body
    FROM questions",
    array () 
);

while ( $row = pg_fetch_array ( $result ) ) {
    $question_body [ $row['question_id'] ] ['body'] = $row['body'];
    $question_index = explode ( " ", $question_body[ $row['question_id'] ] ['body'] ); 
    $question_index = array_unique ( $question_index ); 
}                                                                                                   
var_dump( $question_index );

The problem with this code is that it combines the words in each question. It seems that I cannot use explode, since it seems to make only a single dimensional array.

I also run the following code trying to get question_id unsuccessfully.

#2

while ( $row = pg_fetch_array ( $result ) ) {
    $question_body [ $row['question_id'] ] ['body'] = $row['body'];
    $question_index[ $row['question_id'] ] = explode ( " ", $question_body[ $row['question_id'] ] ['body'] );
    $question_index[ $row['question_id'] ]= array_unique ( $question_index );
}
var_dump( $question_index );

Answer 1

Note that str_replace can accept an array of values as the needle. For example,

$body = str_replace(array(',', '.'), '', $body);

replaces any instance of an element in the array with an empty string.

Alternatively, I'd suggest looking at some pre-built search and indexing libraries. It's a very hard area to get right from scratch and you're much more likely to get reliable results using an already-polished indexing and searching algorithm.

Answer 2

It would be a good idea to first extract all the textual (non-markup) content from the pages by using a DOM parser. See this:

http://stackoverflow.com/questions/795512/how-might-one-go-about-implementing-a-forward-index-in-php

Answer 3

Do yourself a favor and take a look at Zend_Search_Lucene.

Answer 4

I finally decided to use Postgres commands such as LIKE for search. It has been much easier for me to make the data manipulation at DB than at PHP.