Matching the coincidence between the csdn user database and official users and filtering out overlapping users _ PHP Tutorial

Matches the coincidence between the csdn user database and the official user, and filters out overlapping users. Process: 1. the user database obtained from csdn is imported to the local trial editplus. the system prompts that the memory is insufficient and there is no way to solve the problem. my colleague checked it in linux. The basic format is as follows: User name process:
1. obtain the csdn user database and import it locally
After trying editplus for a trial, the system prompts that the memory is insufficient and there is no way to solve the problem. I checked it in my colleague's linux. The basic format is as follows:
Username # Password # Email
Username # Password # Email
Data structure:

The code is as follows:

Create table if not exists 'csdn _ userdb '(
'Id' int (10) not null auto_increment,
'Username' varchar (50) character set gbk not null,
'Password' varchar (50) character set gbk not null,
'Email 'varchar (50) character set gbk not null,
Primary key ('id '),
KEY 'username' ('Username '),
KEY 'email '('email ')
) ENGINE = MyISAM default charset = gbk AUTO_INCREMENT = 1;

It has been suspected that fopen opened files are written to the cache, but practice has proved that the speed is very fast. it should not be written to the cache. the following code is used to import data.

The code is as follows:

$ Link = mysql_connect ('localhost', 'root', 'admin', true );
Mysql_select_db ('csdn ', $ link );
$ Handle = fopen ("C: \ Users \ zhudong \ Desktop \ www.csdn.net. SQL", "r ");
While (! Feof ($ handle )){
$ I ++;
$ Buffer = fgets ($ handle );
List ($ u, $ p, $ e) = explode ("#", $ buffer );
Mysql_query ("insert into csdn_userdb (username, password, email) VALUES ('$ U',' $ P', '$ e')", $ link );
If ($ I % 1000 = 0) echo $ I. "\ n ";
}
Fclose ($ handle );
?>

The above code is very inefficient, so the code after modification is as follows:

The code is as follows:

$ Link = mysql_connect ('localhost', 'root', 'admin', true );
Mysql_select_db ('csdn ', $ link );
$ Handle = fopen ("C: \ Users \ zhudong \ Desktop \ www.csdn.net. SQL", "r ");
$ Perpage = 50;
While (! Feof ($ handle )){
$ I ++;
$ Buffer = fgets ($ handle );
List ($ u, $ p, $ e) = explode ("#", $ buffer );
$ InsertValue [] = "('$ U',' $ P', '$ e ')";
If ($ I % $ perpage = 0) {$ perpage = 0 ){
$ Effectvaluestring = implode (',', $ insertValue );
Mysql_query ("insert into csdn_userdb (username, password, email) VALUES $ effectvaluestring", $ link );
Echo $ I. "\ n ";
$ InsertValue = array ();
}
}
Fclose ($ handle );

In order to find out the factors that affect the efficiency of imported data, we tested the data according to different settings.
Total CSDN user data 6428600
When $ perpage = 500; data imported: 5,902,000; data loss: 526600 loss rate: 8%; data table engine: MyISAM; Index: yes; Total time: 15 minutes
When $ perpage = 200, the total number of imported data: 6,210,200; data loss: 218400; loss rate: 3.3%; data table engine: MYISAM; Index: yes; Total time: 30 minutes
When $ perpage = 200, the total number of imported data: 6,210,200; data loss: 218400; loss rate: 3.3%; data table engine: INNODB; Index: yes; Total time: 65 minutes
When $ perpage = 200, the total number of imported data: 6,210,200; data loss: 218400; loss rate: 3.3%; data table engine: MYISAM; Index: none; total time consumption: 14 minutes (re-create an index after data import)
When $ perpage = 50, the total number of imported data: 6,371,200; data loss: 57400; loss rate: 0.8%; data table engine: MYISAM; Index: None: total time consumed: 20 minutes
The following is a summary based on the above situation:
1. the efficiency of adding an index after the data is imported first is twice as high as that after the data is imported.
2. InnoDB is much less efficient in inserting data from a single process than MYISAM.
3. when perpage is set to 50, the data loss rate is below 1%.

The code is as follows:

Because browser-based execution may cause a timeout problem and the efficiency is low, it may take a lot of time to run through the command line.
At first, I executed the following code:
Php.exe E: \ usr \ www \ importcsdndb. php
However, the following error is reported: call to undefined function mysql_connect
Tossing and finding that php. ini is not loaded
The correct code is:
Php.exe-c E:/usr/local/apache2/php. ini importcsdndb. php

2. import user data that needs to be matched to the local device.
Run the command line to enter msyql (not Baidu)
Run mysql> source C:/Users/zhudong/Desktop/userdb. SQL.
3. compare and filter users
After the comparison program is written, remember to run it in the command line:

The code is as follows:

$ Link = mysql_connect ('localhost', 'root', 'admin', true );
Mysql_select_db ('csdn ', $ link );
$ Handle_username = fopen ("E:/records_username.txt", "");
// $ Handle_email = fopen ("E:/records_email.txt", "");
$ Username_num = $ email_num = $ uid = 0;
While ($ uid <2181106 ){
$ Nextuid = $ uid + 10000;
$ Query = mysql_query ("SELECT * FROM pw_members WHERE uid> '$ uid' AND uid <' $ nextuid '");
While ($ rt = mysql_fetch_array ($ query, MYSQL_ASSOC )){
$ Username = $ rt ['username'];
$ Email = $ rt ['email '];
$ Query2 = mysql_query ("SELECT * FROM scdn_userdb WHERE username = '$ username' OR email =' $ email '");
While ($ rt2 = mysql_fetch_array ($ query2, MYSQL_ASSOC )){
If ($ rt ['password'] = md5 ($ rt2 ['password']) {
If ($ rt2 ['username'] = $ username ){
$ Username_num ++;
Fwrite ($ handle_username, 'own :'. $ rt ['uid']. '| '. $ rt ['username']. '| '. $ rt ['password']. '| '. $ rt ['email ']. 'csdn :'. $ rt2 ['username']. '| '. $ rt2 ['password']. '| '. $ rt2 ['email ']. "\ r \ n ");
Echo 'username _ num = '. $ username_num. "\ r \ n ";
Continue;
}
/*
If ($ rt2 ['email '] = $ email ){
$ Email_num ++;
Fwrite ($ handle_email, 'own :'. $ rt ['uid']. '| '. $ rt ['username']. '| '. $ rt ['password']. '| '. $ rt ['email ']. 'csdn :'. $ rt2 ['username']. '| '. $ rt2 ['password']. '| '. $ rt2 ['email ']. "\ r \ n ");
Echo 'email _ num = '. $ email_num. "\ r \ n ";
}
*/
}
}
Mysql_free_result ($ query2 );
}
$ Uid = $ nextuid;
}
?>

The code you see above is very poor, because it is very inefficient, millions of data, it takes more than 10 hours to run, how can I forget the basics of table connection query? The following is the corrected method.

The code is as follows:

$ Link = mysql_connect ('localhost', 'root', 'admin', true );
Mysql_select_db ('csdn ', $ link );
$ Handle_username = fopen ("E:/records_username.txt", "");
While ($ uid <2181106) {// The number here is the maximum ID of the user library to be compared
$ Nextuid = $ uid + 10000;
$ Query = mysql_query ("SELECT m. uid, m. username, m. password, m. email, u. password as csdn_password, u. email as csdn_email FROM own_members m left join csdn_userdb u USING (username) WHERE m. uid> '$ uid' AND m. uid <= '$ nextuid' AND u. username! = ''");
While ($ rt = mysql_fetch_array ($ query, MYSQL_ASSOC )){
If ($ rt ['password'] = md5 ($ rt ['csdn _ password']) {
$ Username_num ++;
Fwrite ($ handle_username, 'own :'. $ rt ['uid']. '| '. $ rt ['username']. '| '. $ rt ['password']. '| '. $ rt ['email ']. 'csdn :'. $ rt ['username']. '| '. $ rt ['sdn _ password']. '| '. $ rt ['sdn _ email ']. "\ r \ n ");
Echo 'username _ num = '. $ username_num. "\ r \ n ";
}
}
$ Uid = $ nextuid;
Echo 'uid = '. $ uid;
}
?>

The total comparison time is 25 minutes, which is greatly improved compared with the previous 10 hours.
Total name users: 34175
Total membership: 1.7%
1.7% of users with duplicate names are still very serious. I hope this article will help you compare the users on this site.

Tip 1. the user database obtained from csdn is imported to the local trial. when editplus is opened, the system prompts that the memory is insufficient. I cannot find a solution. my colleague checked it in linux. The basic format is as follows: user name...