CN111639267A

CN111639267A - Method for quickly calculating first screen attention posts and related product

Info

Publication number: CN111639267A
Application number: CN202010469664.8A
Authority: CN
Inventors: 郭海萍; 张宇敏
Original assignee: Individual
Current assignee: Individual
Priority date: 2020-05-28
Filing date: 2020-05-28
Publication date: 2020-09-08

Abstract

The invention relates to a method for quickly calculating a first screen attention post and a related product, wherein the method comprises a relationship chain REDIS-ZSE, works REDIS-STRING and a personal work list REDIS-ZSE, and the calculation steps are as follows: firstly, acquiring an attention list by a relationship chain; secondly, obtaining a sequencing attention list from the personal work list; then, the posts with the same number as that required by the first screen are obtained for the concerned objects with the maximum feed ID value of the posts, and the posts of other concerned objects are obtained in batch by a sequentially decreasing algorithm; and finally, merging and sequencing the posts, returning the posts with the corresponding number of the first screen through a work database, and a system and a server applying the method principle. The invention uses the read diffusion model for calculation, reduces the calculation amount of merging and sorting by adopting a descending algorithm, reduces the consumption of calculation resources and the calculation time, accelerates the service processing speed and improves the user experience.

Description

Method for quickly calculating first screen attention posts and related product

Technical Field

The invention relates to the field of computers, in particular to a method for quickly calculating first screen attention posts and a related product.

Background

In the current information age, various social applications and social platforms are widely applied, and people also come from the media age, people publish information through accounts and continuously update contents, each user has an attention list, and all objects in the attention list are attention objects of the user and are attention objects of other people. In practical application, the range of feeds (posts) is wide, and can include various behaviors of friends or objects of interest of a user, after the user logs in the application through a client front end, a lot of feeds need to be displayed by the application, but due to the limitation of the number of displays of a front screen of the client front end, a business process of pulling a friend circle feed flow list is very complicated and has multiple data accesses, a large amount of memory calculation is also needed, network transmission of a large amount of data is realized, the performance is low, the frequency of the feed release of each user is different in the reading of the feed list released by the user, and after the user amount, the data amount and the data of the concurrent amount are gradually increased, the required calculated amount and the calculated time are increased, so that the resource consumption and the calculation time are prolonged, and the required posts cannot be effectively and quickly displayed at the client front end.

Disclosure of Invention

Aiming at the existing defects, the invention provides a method for quickly calculating the first screen attention post and a related product, which can effectively save the calculation time and the calculation resource consumption.

The technical scheme adopted by the invention for solving the technical problems is as follows: a method for quickly calculating a first screen attention post comprises a relationship chain REDIS-ZSE in which all people attention lists and attention relationships are stored, a work REDIS-STRING in which all works issued by all people are stored, and a personal work list REDIS-ZSE in which only key information of the works issued by authors is stored, wherein the calculation steps comprise:

s1, acquiring an attention list of a user through a relationship chain REDIS-ZSTATE;

s2, obtaining the maximum feed ID value of posts of works issued by each attention object in a time reverse order mode after the target time of the attention list through a personal work list REDIS-ZSE, and obtaining a sequencing attention list according to the sequencing of the feed ID value from big to small;

s3, according to the sequencing attention list, posts with the same number as that required by the first screen are obtained for the attention object with the maximum feed ID value, and then posts of other attention objects are obtained in batch by a sequentially decreasing algorithm;

s4, the posts obtained in S3 are merged and sorted in a mode that feed ID values are from large to small, and the posts with the corresponding number in the first screen are returned through the REDIS-STRING according to the data with the required number in the first screen.

Preferably, the sequentially decreasing manner is that the attention object with the largest feed ID value in the sorted attention list acquires the same number of posts as the number required by the first screen according to the number required by the first screen, and the next largest feed ID value acquires the posts one less than the number required by the first screen.

Preferably, when the number of the attention objects in the ranking attention list is greater than or equal to the number required by the first screen, sequentially decreasing until the attention object of one post is obtained in the ranking attention list; and when the number of the attention objects in the sequencing attention list is less than the number required by the first screen, sequentially decreasing until the attention objects with the minimum feed ID value in the sequencing attention list.

Preferably, the feed ID values are arranged in size so that the latest posting time is the largest.

Preferably, the user and the object of interest realize the storage and the acquisition of data among the relationship chain REDIS-ZSE, the works REDIS-STRING and the personal works list REDIS-ZSE through a logic layer.

Preferably, the logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and obtains a self fan list through the relationship chain service, the user inquires and obtains a self attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the works REDIS-STRING is obtained through the work detail service.

Preferably, the focus list is arranged in a size of a focus time of the user on the focus object, and the focus time of the latest focus object is the largest.

A system for quickly calculating a first-screen attention post comprises a relationship chain REDIS-ZSE in which all people attention lists and attention relationships are stored, a work REDIS-STRING in which all works issued by all people are stored, a personal work list REDIS-ZSE in which key information of the works issued by an author is only stored, and a logic layer; the method comprises the steps that a user obtains an attention list of interested objects from a relation chain REDIS-Zset through a logic layer, obtains feed ID values of posts of works issued by each interested object after target time of the attention list in a time reverse order mode from a personal work list REDIS-Zset and sorts the posts in a sorting attention list in a descending mode, obtains posts with the same quantity as that required by a first screen of the interested object with the maximum feed ID value, then obtains posts of other interested objects in batches through a descending algorithm, and the logic layer sorts the obtained posts in a mode that the feed ID values of the posts are reduced from large to small, and returns the posts with the corresponding quantity to the first screen through the works REDIS-STRING according to the data with the quantity required by the first screen.

Preferably, the logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and acquires a vermicelli list of the author through the relationship chain service, the user inquires and calculates and acquires an attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the REDIS-STRING of the works are acquired through a detail service of the works and returned to a head screen.

A server, characterized by: the system comprises a receiving unit, a sorting unit and a sending unit, wherein the receiving unit is used for receiving instruction data sent by a user to an object to be focused through a target application and transmitting the instruction data to the sorting unit; the sequencing unit sequences data in a database according to instructions, the database comprises a relationship chain REDIS-ZSE storing all people attention lists and attention relationships, a work REDIS-STRING storing all works issued by all people, and a personal work list REDIS-ZSE only storing key information of works issued by authors, the sequencing unit acquires and sequences the attention lists of objects concerned by users sequentially through the relationship chain REDIS-ZSE, acquires the maximum feed ID value of posts of works issued by each object concerned after target time of the attention lists through the personal work list REDIS-ZSE, sequences the obtained sequenced attention lists according to the feed ID value from large to small, acquires posts with the same number as that required by a first screen for the object concerned with the maximum feed ID value, and then acquires posts of other objects concerned in batches through a sequentially descending algorithm, and finally, merging and sequencing the acquired posts in a mode that the feed ID value is reduced from large to small, and transmitting data information to a sending unit, wherein the sending unit sends the posts in corresponding quantity to a front-end front-screen page through the REDIS-STRING works according to the data in quantity required by the front screen.

The invention has the beneficial effects that: the method uses the read diffusion model for calculation, the storage structure is simple, each data is correspondingly stored in the corresponding database, only one part of relation data and post data needs to be stored, the storage capacity of the data is small, and a descending algorithm is adopted when the posts are obtained, so that the calculation amount of merging and sorting is reduced, the consumption of calculation resources and the calculation time are reduced, the service processing speed is accelerated, and the user experience is improved.

Drawings

FIG. 1 is a schematic flow diagram of the process of the present invention;

FIG. 2 is a schematic diagram of the system architecture of the present invention;

fig. 3 is a schematic diagram of the server structure of the present invention.

Detailed Description

For the purpose of more clearly illustrating the objects, technical solutions and advantages of the embodiments of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments, for clear and complete description, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without inventive step, are within the scope of the present invention.

Embodiments of the present invention are shown in fig. 1, a method for quickly calculating a first screen attention post includes a relationship chain REDIS-ZSET storing all people attention lists and attention relationships, a work REDIS-STRING storing all works published by all people, a personal work list REDIS-ZSET storing only key information of works published by authors, commonly called feeds, which is content published by users through some applications, the post may include one or more of articles, comments, pictures, videos, change states and other user published contents, the work REDIS-STRING is full data, posts of all attention objects can be obtained through the data, key information of works published by authors includes ID of posts, posting time, posting state and the like, the relationship chain REDIS-ZSET provides attention or paid relationships among all people, each data is correspondingly stored in a corresponding database, the relation data and the post data only need to be stored, the storage capacity of the data is small, and the calculation steps comprise:

s1, obtaining a user attention list through a relation chain REDIS-ZSE, wherein the attention list comprises friends of the user and authors which the user is interested in and actively pays attention to after reading works, the relations between the user and the friends and the authors are stored in the relation chain REDIS-ZSE at the moment, the user attention list is formed when paying attention to a plurality of authors, the user can add or cancel the attention to the authors at any time, in corresponding application, the authors can see their fan lists, namely the attention list, and the user can see their attention list; at this time, the focus list may be arranged in the size of the focus time of the user on the focus object, and the focus time of the latest focus object is the largest, assuming that 3 pieces of data need to be displayed on the front end first screen, as shown in the focus list of table 1, time points T3> T2> T1 at this time

Serial number	Object of interest ID	Time of interest
			1	A3	T3
2	A1	T2
			3	A2	T1

TABLE 1

Wherein: a1, A2, A3 represent different objects of interest, T1, T2, T3 represent points in time of interest when the user focuses on the objects of interest

S2, obtaining the maximum feed ID value of the posts of the work issued by each attention object in a time reverse order mode after the target time through the personal work list REDIS-ZSE, and obtaining the sequencing attention list according to the sorting of the feed ID values from large to small, wherein the feed ID is the unique identification or number of the posts, the target time is the time point that the user needs to acquire the posts, if the user needs to read the posts after the time t1, the works issued by the attention objects in all the attention lists after the time are calculated, the time of the works issued by each attention object is different, when the work is sequenced in the time reverse order mode, for a single attention object, the feed ID values are sized by taking the latest posting time as the maximum, which means that the posting time is more new when the feed ID value is larger, the maximum feed ID value of each attention object is obtained through calculation, then the maximum feed ID values of different attention objects are compared and are arranged in descending order to obtain a sequencing attention list, so that the first new post is equal to the post with the first time value, namely the post with the maximum feed ID value, the second new post is the post with the second time value, and the rest can be analogized to obtain the post list sequenced according to the time value or the feed ID value, as shown in the sequencing attention list of Table 2, the feed ID value F99 is more than F98 is more than F97 is more than …,

serial number	Object of interest ID	Feed ID after F
			1	A1	F99、F96、F93、F90
2	A3	F98、F95、F92、F89
			3	A2	F97、F94、F91、F88

TABLE 2

Wherein, A1, A2 and A3 represent different objects of interest, F represents a target time point, and F99, F98 and F97 represent the maximum feed ID values of the different objects of interest after F

From Table 2, it can be determined that the "first new post", i.e., the post with the largest feed ID, was posted by the person with sequence number 1 in the table, i.e., the first person listed in the table, i.e., person A1, the second new post was posted by the person with

sequence number

1 or 2 in the table, i.e., A1 or A3, and the third new post was posted by the person with

sequence number

1, 2 or 3 in the table, i.e., A1 or A3 or A2;

s3, according to the ordering concern list, the concern object with the largest feed ID value is obtained the same number of posts as the first screen, then the posts of other concern objects are obtained in batch by the descending algorithm, such as the descending relation list in Table 3,

serial number	Object of interest ID	Feed ID after F
			1	A1	F99、F96、F93
2	A3	F98、F95
			3	A2	F97

TABLE 3

Because the number, time and the like of the bouquets of each attention object are different, the data volume to be calculated is reduced in such a way, but the number of posts which need to be displayed on the first screen of the front end of a client and meet the requirements cannot be omitted, the sequential decreasing way is that the attention object with the largest feed ID value is used for acquiring posts with the same number as the required number of the first screen in the sequencing attention list according to the required number of the first screen, the next largest feed ID value is used for acquiring the posts in a way of being less than the required number of the first screen, at the moment, corresponding to the table 2, if a first new post, a second new post and a third new post are acquired, namely, the post with the first largest time value, the post with the second largest time value and the post with the third largest time value are acquired, 3 posts are acquired in

number

1, 2 are acquired in

number

2, 1 is acquired in number 3, and so on, if the posts with the first largest time value are acquired, "post with the second largest time value" …. "post with the nth largest time value", on the premise that N attention objects are corresponded, the first person needs to take N, the second person needs to take N-1, …., and the nth person needs to take 1; when the number of the attention objects in the attention list is different from the number required to be displayed on the first screen, and the number of the attention objects in the sequencing attention list is larger than or equal to the number required by the first screen, sequentially decreasing until the attention object of one post is obtained in the sequencing attention list; and when the number of the attention objects in the sequencing attention list is less than the number required by the first screen, sequentially decreasing until the attention objects with the minimum feed ID value in the sequencing attention list. At this time, M1 represents the number of to-be-merged-sorted items in the conventional method, M2 represents the number of to-be-merged-sorted items in the method, P represents the number of attention objects in the attention list, N represents the number that needs to be displayed on the front screen of the client, and when P is 1 in the attention list, the same numbers of M2 and M1 are both N; when N is 1, M2 and M1 are both 1 in the same way; when P >1 and N >1, the comparison of the two is as follows,

1) in the method, the last M2 is obtained in a descending way, and belongs to an arithmetic sequence, the sum of arithmetic sequence is (first term + last term) x (term/2),

2) if P > ═ N, then M2 ═ N + (N-1) + … +1 ═ N (N + 1)/2; otherwise (P < N), M2 ═ N + (N-1) + … + (N-P) ═ N-P/2 ═ P ═ N-P/2

3) M1 ═ N × min (P, N) ═ min (N × N, P × N), that is, if P > ═ N, M1 ═ N × N; otherwise (P < N), M1 ═ P × N

4) When P > -N, M2 ═ N (N +1)/2< M1 ═ N, mathematical induction:

A. assuming that N is 2, M2 is 3, and M1 is 4, this formula holds true.

B. Assuming that N is K, the formula holds, i.e., K (K +1)/2< K

C. When false N is K + 1:

K*(K+1)/2+(K+1)<＝K*K+(K+1)

K*(K+1)/2+2(K+1)/2<＝K*K+K+1

(K+2)*(K+1)/2<＝K*K+K+1<＝K*K+2K+1

(K+1)((K+1)+1)/2<＝K*K+K+1<＝(K+1)*(K+1)

(K+1)((K+1)+1)/2<＝(K+1)*(K+1)

since N is also true when N +1 is stated, it is concluded that M2 is N (N +1)/2< M1 is N for an arbitrary natural number N, i.e., M2< M1

5) When P < N, M2 ═ P × N-P/2 < M1 ═ P × N.

Therefore, it can be concluded that, for any natural number N and P, M2< ═ M1 exists, and M2< M1 in the case that the relationship chain list P >1 and the number of page display pieces N >1 in a common relationship chain scene, which fully explains that the method effectively reduces the calculation amount of merging and sorting, reduces the consumption of calculation resources and the calculation time, accelerates the service processing speed, and improves the user experience.

S4, merging and sorting the posts obtained in the step S3 in a mode that feed ID values are from large to small, and returning the posts with the corresponding number to the first screen through the REDIS-STRING according to the data with the number required by the first screen, wherein the result of the merged and sorted data is 'F99, F98, F97, F96, F95 and F93', and the result of the front screen value requiring 3 pieces of data is 'F99, F98 and F97', so that the calculation amount and the calculation resources of merging and sorting are greatly reduced.

In a further improvement, as shown in fig. 2, the user and the interested object realize the storage and the acquisition of data among the relationship chain REDIS-ZSET, the work REDIS-STRING and the personal work list REDIS-ZSET through a logic layer, so that the logic structure of the database data is provided, the extraction and the storage of each data are effectively realized, and the logic layer can be realized by adopting any existing data model. The logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in a work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and obtains a self fan list through the relationship chain service, the user inquires and obtains a self attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the works REDIS-STRING is obtained through the work detail service.

According to the method for quickly calculating the first screen attention post, a system is formed, as shown in fig. 2, the system for quickly calculating the first screen attention post comprises a relationship chain REDIS-ZSE in which all people attention lists and attention relationships are stored, works REDIS-STRING in which all works issued by all people are stored, a personal work list REDIS-ZSE in which only key information of the works issued by authors is stored, and a logic layer; the method comprises the steps that a user obtains an attention list of an attention object from a relation chain REDIS-Zset through a logic layer, obtains feed ID values of posts of works issued by each attention object after target time of the attention list in a time reverse order mode from a personal work list REDIS-Zset and sorts the posts in a sorting attention list in a descending mode, obtains posts with the same quantity as that required by a first screen of the attention object with the maximum feed ID value, then obtains posts of other attention objects in batches through a descending algorithm, and the logic layer sorts the obtained posts in a merging mode in a feed ID value descending mode and returns the posts with the corresponding quantity to the first screen through the works REDIS-STRING according to the data with the quantity required by the first screen. The logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in a work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and acquires a vermicelli list of the author through the relationship chain service, the user inquires and calculates and acquires an attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the REDIS-STRING of the works are acquired through a detail service of the works and returned to a head screen.

A server, as shown in fig. 3, includes a receiving unit 1, a sorting unit 2, and a sending unit 3, where the receiving unit 1 is configured to receive instruction data sent by a user through a target application to an object to be focused and transmit the instruction data to the sorting unit 2; the sorting unit 2 sorts data in a database 4 according to instructions, the database 4 comprises a relationship chain REDIS-ZSE storing a list of all people's concerns and concerns, a works REDIS-STRING storing all works issued by all people, and a personal works list REDIS-ZSE storing key information of works issued by authors, the sorting unit 2 sequentially acquires and sorts the concerns of the objects of interest of users through the relationship chain REDIS-ZSE, acquires the maximum feed ID value of posts arranged in a time reverse manner of the works issued by each object of interest after the target time of the concerns list through the personal works list REDIS-ZSE, sorts and acquires the sorted concerns list according to the feed ID value from large to small, acquires posts with the same number as that required by a first screen for the object of interest with the maximum feed ID value, and then acquires posts of other objects of interest in batch by a sequential descending algorithm, and finally, merging and sequencing the acquired posts in a mode that feed ID values are reduced from large to small, and transmitting data information to a sending unit 3, wherein the sending unit 3 sends the posts with corresponding quantity to a front-end front-screen page through a REDIS-STRING work according to the data with the quantity required by the front-screen.

It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims

1. A method for quickly calculating a first screen attention post is characterized by comprising the following steps: the method comprises a relationship chain REDIS-ZSE in which an attention list of all people and attention relationships are stored, a work REDIS-STRING in which all works issued by all people are stored, and a personal work list REDIS-ZSE in which key information of the works issued by authors is only stored, wherein the calculation steps comprise:

s2, obtaining the maximum feed ID value of posts of works issued by each attention object in a time reverse order mode after the target time of the attention list through a personal work list REDIS-ZSE, and obtaining a sequencing attention list according to the sorting of the feed ID values from big to small;

2. The method for rapidly calculating the first screen attention post of claim 1, wherein: the sequential decreasing mode is that according to the number required by the first screen, the attention object with the largest feed ID value is used for acquiring the posts with the same number as the number required by the first screen in the sequencing attention list, and the post with the second largest feed ID value is acquired in a mode that the number of the posts is less than the number required by the first screen.

3. The method for rapidly calculating the first screen attention post of claim 2, wherein: when the number of the attention objects in the sequencing attention list is larger than or equal to the number required by the first screen, sequentially decreasing until the attention object of one post is obtained in the sequencing attention list; and when the number of the attention objects in the sequencing attention list is less than the number required by the first screen, sequentially decreasing until the attention objects with the minimum feed ID value in the sequencing attention list.

4. The method for rapidly calculating the first screen attention post of claim 1, wherein: the feed ID values are arranged in size so that the latest posting time is the largest.

5. The method for rapidly calculating the first screen attention post of claim 1, wherein: the user and the attention object realize the storage and the acquisition of data among a relationship chain REDIS-ZST, a work REDIS-STRING and a personal work list REDIS-ZST through a logic layer.

6. The method for rapidly calculating the first screen attention post of claim 5, wherein: the logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in a work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and obtains a self fan list through the relationship chain service, the user inquires and obtains a self attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the works REDIS-STRING is obtained through the work detail service.

7. The method for rapidly calculating the first screen attention post of claim 1, wherein: the focus list is arranged by the focus time of the user on the focus object, and the focus time of the latest focus object is the maximum.

8. A system for rapidly computing first screen attention posts, comprising: the system comprises a relationship chain REDIS-ZSE in which an attention list of all people and attention relationships are stored, a work REDIS-STRING in which all works issued by all people are stored, a personal work list REDIS-ZSE in which key information of the works issued by an author is only stored, and a logic layer; the method comprises the steps that a user obtains an attention list from a relation chain REDIS-Zset through a logic layer, obtains feed ID values of posts of works issued by each attention object in a time reverse order mode after target time of the attention list from a personal work list REDIS-Zset and sorts the posts in a sorting attention list from big to small mode, obtains posts with the same quantity as that required by a first screen of the attention object with the maximum feed ID value, then obtains posts of other attention objects in batches through a descending algorithm, and the logic layer sorts the obtained posts in a feed ID value from big to small mode and returns the posts with the corresponding quantity to the first screen through the REDIS-STRRIP according to data with the quantity required by the first screen.

9. The system for rapidly calculating the first screen attention post of claim 8, wherein: the logic layer comprises a relation chain service, a work detail service, a publisher service and a personal homepage service, wherein an author publishes a post through the publisher service, the post is stored in a work REDIS-STRING and displayed in a personal homepage through the personal homepage service, and key information of the post is stored in a personal work list REDIS-ZSE; the user actively pays attention to the author after reading the works, the relationship between the user and the author is stored in a relationship chain REDIS-ZSE through a relationship chain service, the author inquires and acquires a vermicelli list of the author through the relationship chain service, the user inquires and calculates and acquires an attention list and a personal work list REDIS-ZSE through the relationship chain service, and data of the REDIS-STRING of the works are acquired through a detail service of the works and returned to a head screen.

10. A server, characterized by: the system comprises a receiving unit, a sorting unit and a sending unit, wherein the receiving unit is used for receiving instruction data sent by a user to an object to be focused through a target application and transmitting the instruction data to the sorting unit; the sorting unit sorts data in a database according to instructions, the database comprises a relationship chain REDIS-ZSE storing all people attention lists and attention relationships, a work REDIS-STRING storing all works issued by all people, and a personal work list REDIS-ZSE storing key information of works issued by authors, the sorting unit acquires and sorts the attention lists of users sequentially through the relationship chain REDIS-ZSE, acquires the maximum feed ID value of posts of works issued by each attention object in a time reverse order after target time through the personal work list REDIS-ZSE, sorts and acquires the sorted attention lists according to the feed ID value from large to small, acquires posts with the same number as that required by a head screen for the attention object with the maximum feed ID value, and then acquires posts of other attention objects in batch through a sequentially descending algorithm, and finally, merging and sequencing the acquired posts in a mode that the feed ID value is reduced from large to small, and transmitting data information to a sending unit, wherein the sending unit sends the posts in corresponding quantity to a front-end front-screen page through the REDIS-STRING works according to the data in quantity required by the front screen.