CN101799830A - Flow data processing method capable of realizing multi-dimensional free analysis - Google Patents

Abstract

The invention provides a flow data processing method capable of realizing multi-dimensional free analysis. OLAP (online analytical processing) technology is adopted to correspondingly classify original flow data and establish relevant association. Simultaneously, relevant association is also established between each classification of flow data and all dimensions, and all classified data share dimension data, so the multi-dimensional free analysis on the flow data can be realized by using multi-dimension to analyze the data layer by layer. A user can take any part of the flow data out in a report form, perform gradual free analysis by using a plurality of the dimensions and select the flow data continuously until a final desired data report is obtained.

Description

Can realize the flow data processing method of multi-dimensional free analysis

Technical field

The present invention relates to the web traffic analysis, relate in particular to a kind of flow data processing method of realizing multi-dimensional free analysis.

Background technology

Web traffic analytic system commonly used at present, the more built-in form commonly used in capital, province form for example, search engine form, keyword form etc., to be each form get final product with regard to only carrying out the polymerization computing to database table wherein for the dimension of these composition data flows, the data processing method that it adopted, and do not have related between form and the form, so these forms all are " static state ", that is to say that the client takes after these forms, he can't carry out other operations.Even he produces query to the partial data in the form, what also has no idea to do, because all analyze dimensions, all at whole datas on flows, rather than at wherein a part of data on flows of certain form.Seemed to provide the dimension of many analyses, but all isolate between the dimension, all dimensions all are based on whole datas on flows (certainly, here said whole data on flows, say accurately, should be the whole datas on flows in a certain period) analyze, at this moment, the client can only see similar three isolated forms shown in Figure 1.The client takes after this form, a superficial understanding also can only be arranged the flow of oneself, if the client has more senior demand, think the alternate analysis form, want the partial data in form is wherein analyzed separately, for example, the client want to check from " Beijing " and " Guangdong " and be in the flow of coming by the Google search, which keyword these visitors have searched for, and which commodity each keyword all bought at last, and this class instrument just can't solve so.

Summary of the invention

At above defective, the purpose of this invention is to provide a kind of flow data processing method, can realize the multi-dimensional free analysis of data on flows.

For achieving the above object, the present invention is achieved through the following technical solutions:

A kind of flow data processing method of realizing multi-dimensional free analysis adopts OLAP (onlineanalytical processing on-line analytical processing) technology, may further comprise the steps:

(1), original data on flows is carried out standard and also be divided into 6 classifications accordingly, be respectively page access data, session access data, guest access data, ad data, electronic commerce data and click data, for these several classifications are set up OLAP fact table (FactTable) respectively, set up corresponding related between the different flow data by Session ID and visitor's identification identifier;

(2), set up suitable dimension, similar dimension is grouped into a class, sets up corresponding dimension table, data generate unique major key in the table, carry out related with the data on flows session of above-mentioned 6 big classifications respectively then at each data category;

(3), set up the OLAP data cube according to above-mentioned fact table and dimension table, by MDX (Multidimensional Expressions Multidimensional Expressions) the final form of language generation.

The present invention simultaneously, between the data on flows and all dimensions for each classification, has also set up corresponding association by original data on flows being sorted out accordingly and being set up corresponding association.All shared these dimension datas of data after all are sorted out, therefore can use various dimensions to go profile data successively, realized multi-dimensional free analysis to data on flows, the user can take out any a part of data on flows in the form, use a plurality of dimensions progressively to analyze freely, constantly screening is until the data sheet of finally being wanted.

Description of drawings

The present invention is described in further detail according to embodiment and accompanying drawing below.

Fig. 1 is the isolated mutually report form synoptic diagram that existing web traffic analytic system is generated;

Fig. 2 is the correlationship figure that adopts between the method for the invention classification data afterwards;

Fig. 3 is the data on flows of session data in the example and the graph of a relation between the dimension;

Fig. 4-Fig. 6 sets up dimension table and is generated the schematic diagram of fact table by dimension table at raw data;

Fig. 7 is the OLAP data cube of setting up according to fact table and dimension table.

Embodiment

In order to realize the analysis data on flows of multi-dimensional free, our WebDissector background system has adopted OLAP (online analytical processing on-line analytical processing) technology, original data on flows is carried out standard and is divided into 6 classifications accordingly, be respectively page access data, session access data, guest access data, ad data, electronic commerce data and click data, data after sorting out are not what isolate, but certain association is arranged, graph of a relation is as shown in Figure 2.Association between them, be to realize by the identifier of two cores, one of them is a Session ID, is used for the data with a session (we will sum up in the point that in the session with a close page access of visitor), comprise ad data, click data etc. are gone here and there together; Another one is visitor's identification identifier, is used for unique sign visitor, we with this identifier with visitor visit data string for a long time to together.Data on flows after all are sorted out has all comprised this two important indications, therefore could set up the relation of the various complexity among the last figure.

Setting up suitable dimension at each data category, between the data on flows and all dimensions for each classification, also is that corresponding association is arranged; The act session data is an example, as shown in Figure 3.

As shown in Figure 4, provided the raw data of a session, wherein province and city are of a sort, it is classified as " geographical environment table " as a dimension, set up corresponding dimension table, search engine and keyword are classes, it are included in " session source table " as a dimension, set up corresponding dimension table, and be the major key of data allocations in each dimension table; The major key of respective dimension kilsyth basalt is set up corresponding OLAP fact table (FactTable) by reference.Promptly organize relation between the data, so just produced structure as shown in Figure 5 by setting up fact table and dimension table.

In this way, the memory complexity of background system just greatly reduces, the consistance of data is guaranteed, simultaneously, major key in the respective dimension kilsyth basalt obtains data on flows and generates corresponding form by reference, can reuse identical dimension table between the different fact tables, for example the data of ecommerce also can be quoted same geographical dimension kilsyth basalt and source dimension table, as shown in Figure 6.

Because a fact table direct correlation several dimension tables, we just can set up data warehouse memory structure as shown in Figure 7, and it is one typical 3 data cube of tieing up, each cell all correspondence a batch session.The client can precisely orient the flow that he pays close attention to according to the wish of oneself, for example among Fig. 7 shown in the dark cell, be " from Beijing " and the flow of coming the website by " Sogou search engine " search " popularization " keyword, always have 37 visits, by setting up this OLAP data cube, further can utilize MDX (Multidimensional Expressions Multidimensional Expressions) the final form of language generation.

What above-mentioned example was showed is the parsing process of 3 dimensions, we in addition can take out these " flow cell lattice ", as a new data cube, use new dimension to analyze, finally can reach and adopt various dimensions to go profile data successively, constantly screening is until the data sheet of finally being wanted.

Claims (1)

1. the flow data processing method that can realize multi-dimensional free analysis adopts OLAP (onlineanalytical processing on-line analytical processing) technology, it is characterized in that may further comprise the steps:

(1), original data on flows is carried out standard and also be divided into 6 classifications accordingly, be respectively page access data, session access data, guest access data, ad data, electronic commerce data and click data, for these several classifications are set up OLAP fact table (FactTable) respectively, set up corresponding related between the different flow data by Session ID and visitor's identification identifier;

(2), set up suitable dimension, similar dimension is grouped into a class, sets up corresponding dimension table, data generate unique major key in the table, carry out related with the data on flows session of above-mentioned 6 big classifications respectively then at each data category;

(3), set up the OLAP data cube according to above-mentioned fact table and dimension table, by MDX (Multidimensional Expressions Multidimensional Expressions) the final form of language generation.

Images