Php+mysql simple implementation of Chinese word segmentation full-text indexing

Intermediary transaction http://www.aliyun.com/zixun/aggregation/6858.html ">seo diagnose Taobao guest cloud host technology Hall

I believe many people have studied the Chinese full-text indexing MySQL problem, the current MySQL does not support Chinese full-text indexing, helpless ~ ~ ~

On the Internet to find more n on the php+mysql Chinese Full-text index articles are not a good, for the general rookie do not know how to use, PHP Chinese participle has ddede, Dede Thesaurus is not very powerful, but also calculate can use it!

There is also a "scws-simple Chinese word segmentation system," the current for the general rookie do not know how to use, although someone compiled the Windows version, but the installation is still very troublesome, I have not tested!

//==========================

Nonsense don't say much, first go to dede download word algorithm function and thesaurus; Download address: http://www.dedecms.com/html/chanpinxiazai/20061229/3.html

There is a detailed use of the instructions, I am here to cite an example of use;

MySQL Table name: Music

Field: Title,tag

Require ("lib_splitword_full.php");

$str = "Here is the content you want to participle, generally do not exceed 1024KB will be better, otherwise it will be very slow!"

$SP = new Splitword ()//instantiated

$DD = Explode ("", $sp->splitrmm ($STR));

$i = 0;

foreach ($dd as $key => $var) {

if (strlen ($var) >2)//UTF8 encoded please set to 3, because UTF8 encoded in Chinese is generally more than 3 bytes, filter single words do not save!

{

$str. =base64_encode ($var). "//Because MySQL does not support Chinese full-text indexing, we have to convert words into letters or numbers in participle, where I choose Base64 encoding

Of course, we can also convert Chinese into region code, but I tested the method of the region code, feel very general, so I recommend you to use Base64

$i + +;

}

if ($i >=50) break;//Here is set how many phrases you want to store, if the article may have n many phrases, you can set a larger point, but if you are on the title participle, 50 is enough!

This operation is not coded results: the content of this word is generally not more than 1024KB a point otherwise

Base64 encoded results: 1ela7w== t9a0yg== xnri3q== 0ruw4w== srvsqg== s6y5/q== mtaynetc, 0ru14w== t/hu8g==

}

?>

So we can divide the word and convert it into encoded word into the tag field, of course the tag field must be set to Full-text index full text and 6184.html "> Datasheet type must be MyISAM."

The tag field data type is char and varchar, text, about how to create a MySQL full-text index, you can search the MySQL creation index everywhere!

The following example is a Full-text indexing query based on the contents of the tag field:

Include ("Mysql.class.php");

$rs = $DB->get_one ("Select Title,tag from Music where id= $id")//Read title and tag field contents by ID

$title = $rs [' title '];

$sql = $DB->query ("select Title,match (tag) against ('. Trim ($rs [' tag '])." In BOOLEAN MODE as score from music where MATCH (tag) against (' ". Trim ($rs [' tag '])." In BOOLEAN MODE) (score DESC limit 21 ");//This is the query 20 similar content, and according to the similar sort, 21 is because the same record will match!

while ($rs = $DB->fetch_array ($sql))

{

if ($rs [' title ']!= $title)

{

echo $rs [' title '].
";

}

}

$DB->close ();

?>

If you want to see the effect can go to my station to see the demo: http://www.dj965.com

The above example realizes the Php+mysql simple Chinese word segmentation full-text Index!

Hope for the same as me rookie have some help!

I blog:http://hi.baidu.com/web20/