PHP enables e-commerce orders to automatically confirm the receipt of Redis queue method

This article mainly introduces the PHP implementation of e-commerce orders to automatically confirm the receipt of Redis queue method, interested in a friend's reference, I hope to be helpful to everyone.

First, the scene

Before the e-commerce platform, the user in the receipt of goods, most of them will not take the initiative to confirm the receipt, resulting in the business of the payment, the merchant various complaints, so according to demand, to do an order after the delivery of X days automatically confirm receipt. The so-called automatic confirmation of receipt of orders, that is, at a specific time, execute an UPDATE statement, change the status of the order.

Second, the idea

The most cumbersome approach, through the Linux Background Timer task, query the eligible orders, and then update. Ideally, this would be fine if there were orders to update every minute. The platform is too small, and most of the seller's delivery time is also dense, not scattered in 24 hours per minute. Then, the timing of the task, the query too much, not suitable. Here you can automatically confirm the receipt of the order information stored on other media, such as REDIS,MEMCACHE,RABBITMQ, and then execute the script from the previous media to obtain the order information to judge, here can greatly reduce the database query pressure.

The producer of the Redis queue

In this regard, we choose every day at two o'clock in the morning, through the Linux scheduled task to confirm the receipt of the order information to check out, and then stored on the Redis, Redis on the queue we choose, the characteristics of the queue processing is FIFO, the front of the data in the query order, the delivery time by the order, So the first out of the queue is definitely the nearest order from the specified automatic receipt time. The code is as follows

$successCount =0; $failCount =0; $screen _time = 3600*24*9;//set filter days $data = Array (), $now _time = time (),//query meets the required data $sql= " Select Id,send_time as Deliver_time from ' order ' where is_send=1 and Is_del=0 and Is_cancel=0 and Is_token=0 and Send_time >0 and Send_time + {$screen _time} < $now _timeorder by send_time asc "; $res = $con->query ($sql);//record and purge data when the queue has data while ($redis->llen (' Auto_recevice_order ')) {$txt = ' execution time: '. Date (' y-m-d h:i:s '). ', Info: '. $redis->rpop (' Auto_ Recevice_order '); File_put_contents ('./autotoken/fail_log.txt ', $txt. " \ r \ n ". Php_eol,file_append); $failCount + +;} Repopulate the data into the queue while ($row = $res->fetch_assoc ()) {$successCount + +; $redis->lpush (' Auto_recevice_order ', json_ Encode ($row 1));} $con->close (); $success =date (' y-m-d h:i:s '). ': [Push success]: This successful push data: '. $successCount. ', record last processing failed data: '. $failCount. ' Article \ r \ n "; File_put_contents ('./success_log.txt ', $success. " \ r \ n ". Php_eol,file_append);

Customers of Redis queues

Queue of consumers do not through the timing of the Linux task to do, with the Linux screen+php CLI mode to execute PHP script, the consumer only need to continuously read order information from the queue, and then determine the order information in the delivery time, if the requirements of automatic receipt, The UPDATE statement is executed. At the same time, if the receipt is not reached, and with the receipt of the time span is relatively large, you can let the PHP script sleep a certain amount of time, the amount of time to adjust the design, to get out of the time required for the order to be re-pushed to the Redis queue, but also the top of the queue. To get it next time. The code is as follows:

$set _time = 3600*24*10;//set a few days after the automatic receipt while (true) {if ($i%30==0) {usleep (10);//prevents the while loop from excessively high CPU usage}if ($redis->llen (' Auto_recevice_order ') {$data = Json_decode ($redis->rpop (' Auto_recevice_order ')), $id = (int) $data->id;// Convert data to shaping $deliver_time = (int) $data->deliver_time;//convert data to shaping $res1 = $res 2 =false; $now _time = time (); if ($deliver _ time+ $set _time) < $now _time) {$sql 1 = "Update ' order ' set ' is_token ' = ' 1 ', ' token_time ' = $now _time where id= $id and Is_sen D=1 and Is_del=0 and Is_cancel=0 and is_token=0 and Send_time + {$set _time} < $now _time "; $res 1 = $con->query ($sql 1);//Update Data $rows = Mysqli_affected_rows ($con), if ($rows) {$ip = $this->getip (); $sql 2 = " Insert INTO ' order_log ' (' order_id ', ' log_msg ', ' log_ip ', ' log_role ', ' log_user ', ' log_order_state ', ' Log_time ') VALUES ( $id, ' System automatic receipt ', ' $ip ', ' System ', ' server ', ' receipt ', $now _time) ';//write the Order Log $res 2 = $con->query ($sql 2);//Add log Data}} if ($res 1==false) { Reinsert data that is not up to the condition into the queue $redis->rpush (' Auto_recevice_order ', Json_encode (' id ' = ' = ' $id, ' deliver_time '= $deliver _time))); }} $i + +;}

Executing PHP scripts here requires the use of the Linux screen or supervisor, Nohup daemon. The specific use of Baidu can be self-contained in the same script it is best to have the necessary log records.

Third, thinking

As the business grows, in the same second in the queue, there are many orders that need to be processed, and one can only fetch a relevant order information from the queue at a time, the model of multiple consumers may be used, in which case the lock mechanism can be used to ensure that a message can reach only one consumer. When the Redis data reaches a certain amount, it can also adjust the producer's execution frequency and the corresponding condition appropriately.

This PHP implementation of e-commerce orders automatically confirm receipt of Redis queue is a small part of the whole content to share to everyone, I hope to give you a reference, but also hope that we support the script home.

Related recommendations:

E-commerce website automatically confirms how the receipt is realized

Php method for obtaining 6-bit random numbers that do not exist in Redis

PHP implementation of Redis Message Queuing to publish micro-blogging methods