WO2009004231A2

WO2009004231A2 - Information sorting method

Info

Publication number: WO2009004231A2
Application number: PCT/FR2008/051064
Authority: WO
Inventors: Romuald Poirot; Nicolas Sede
Original assignee: Compario
Priority date: 2007-06-13
Filing date: 2008-06-13
Publication date: 2009-01-08
Also published as: WO2009004231A3; FR2917518A1; FR2917518B1

Abstract

The invention relates to a method for determining the result of the sorting process of information relative to a plurality of entities, according to at least one sorting value associated with at least one sorting criterion. The sorting criterion is selected from a plurality of sorting criteria each relating to a characteristic of said entities and having one or more sorting criterion values, said sorting value being selected from one or more sorting values of said sorting criterion. The method is characterised in that: each sorting criterion is associated with a dimension of a multi-dimensional database (11), said base being built from data relating to said entities, each of the sorting values of a sorting criterion being associated with a member of the dimension associated with said sorting criterion; and the result of the sorting process is determined based on the content of the multi-dimensional database (11), the coordinates of which are the member associated to the sorting value on the dimension corresponding to said sorting criterion.

Description

"Information sorting process"

The present invention relates to an information sorting method. It also aims a system implementing this method. The invention relates more particularly to a method of sorting information in an electronic catalog according to sorting criteria relating to information relating to the entities to be sorted.

Currently, many websites offer tools for sorting information of products offered for sale in electronic catalogs, with the objective of providing product comparators. These sorting tools allow users to sort the products of the same category according to several of the sorting criteria proposed, such as for example price, brand, color, etc. For each of the sorting criteria proposed, several sorting values are available, such as, for example, price bands when the sorting criterion is the price.

The sorting methods implemented in the sorting tools currently proposed make it possible to sort information in several steps. At first, the user selects one or more sort values for one or more sorting criteria. In a second step, using a query generated according to the selected sorting criteria values, a catalog database containing all the products concerned by the sorting is accessed and each sorting value of each criterion sorting is tested on each of the products involved in sorting. Finally, the result of the sort is displayed. However, these sorting methods have several disadvantages. Indeed, it happens quite frequently that a user results in a zero result for a combination of sorting values of the sorting criteria he has selected. In addition, these sorting processes require very large computer resources and the time to perform a sorting increases with the frequentation of the site.

Other sorting methods, incorporating a pre calculation of the proposed combinations of sorting values, make it possible to avoid obtaining a zero sorting result. However, these sorting methods pose volumetric problems and have sorting capabilities limited to sorting combinations composed of a dozen sorting values and / or sorting criteria.

An object of the invention is to provide a method and a system for determining the result of a sort of information faster, requiring less computing resources than current sorting processes.

Another object of the invention is to provide a method and a sorting system having a greater sorting capacity.

The present invention makes it possible to achieve the aforementioned aims by a method for determining the result of a sort of information relating to a plurality of entities, according to at least one sorting value of at least one sorting criterion: the criterion sorting agent being selected from a plurality of sorting criteria each relating to a characteristic of each of the entities, and having one or more sorting criterion values, and - said sorting value being selected from one or more sorting values of the sorting criterion. sorting; characterized in that

each sorting criterion is associated with a dimension of a multidimensional database, this base being constructed from the characteristics of the entities, each of the sorting values of a sorting criterion being associated with a member of the dimension associated with this sorting criterion, and the sorting result is determined according to the content of the multidimensional database whose coordinates are the member associated with the sorting value on the dimension corresponding to the sorting criterion.

By sorting information relating to a plurality of entities, we mean a computer selection of entities satisfying sorting values of selected sorting criteria. The result of such a selection can be the number of entities satisfying the sorting values, the entity identifier or any other data relating to the entities concerned.

The method according to the invention makes it possible, by the use of a multidimensional database as described above, to perform sorting and to determine the results of sorting performed more rapidly, more flexible, less expensive and requiring less computing resources when sorting.

The multidimensional database is constructed based on the characteristics of the entities on which the sorting of information and the proposed values and sorting criteria is carried out. Thus, each content of the multidimensional database corresponds to the result of the sorting carried out with the combination of sorting values associated with the members, the coordinates of this content, for the sorting criteria associated with the dimensions on which these members are located. In other words, the result of a sorting performed for one or more sort values selected for one or more sorting criteria is determined according to the content of the multidimensional database whose coordinates are the members associated with each of said sorting values. on all dimensions, associated with each of said sorting criteria.

The sorting criteria and sorting values for each sorting criterion can be determined, modified, deleted or added by an administrator, using a back office tool, based on the entity data.

In a particular embodiment, said expert, the sorting values for each sorting criterion can be displayed in a drop-down menu associated with the sorting criterion. In this embodiment the user can make a multi selection of sort values. In another embodiment, called "drilldown", the sort values for each sort criterion can be displayed in the form of a tree.

In the particular embodiment where the sorting result can include the number of entities satisfying the sorting values for the sorting criteria. In this case, each content of the multidimensional database corresponds to the number of entities satisfying the combination of sorting values associated with the members, coordinates of this content, for the sorting criteria associated with the dimensions on which these members are located. When the combination of selected sort values results in several different coordinate contents in the multidimensional database, the result of sorting is the sum of these contents. Advantageously, the method according to the invention may comprise, for a preceding sorting carried out according to at least one preceding sorting value, a determination of at least one sorting value, said to be available, for at least one other sorting criterion, which may be combined with said previous sorting value, to perform a new sort giving a non-zero result. The sort values giving a zero sorting result in combination with at least one previous sorting value are not proposed so that no each sorting performed results in a zero result. Thus, the method according to the invention makes it possible to update the sorting values for each sorting criterion so as to avoid a combination of sorting values resulting in a zero result. In other words, whatever the combination of sort values selected, there will be at least one product satisfying this combination of sort values. The method according to the invention makes it possible, on the one hand, to avoid the frustration created in a user by zero result sorting and, on the other hand, to save the time lost for sorting zero results.

In addition, the method according to the invention may further comprise a display of the result of a new sort for at least one sorting value available. Indeed, when a sorting value is available to be combined, in the framework of a future sorting, with at least one previous sorting value used for a previous sorting, the method according to the invention makes it possible to display the result sorting to come.

The multidimensional database is accessed by a Multi Dimensional Expression (MDX) query generated by an MDX query builder from the combination of sort values of selected sort criteria.

Advantageously, the multidimensional database can be constructed from a so-called intermediate database, comprising for each entity an identifier and a sorting value for each sorting criterion. According to a first embodiment, such a database can be organized in the form of a two-dimensional table, for each category of entities, containing a column for the product identifier and a column for sorting criteria. According to a second embodiment, the intermediate database can be organized in the form of a table by sorting criterion. This second embodiment offers better performance when feeding the multidimensional base from the intermediate base.

Such an intermediate database makes it possible to update the multidimensional database simply, quickly and without requiring significant computer resources.

The intermediate database can be constructed from a database, called a catalog, using SQL type queries to access the contents of the catalog database. For each entity of a feature category, a sort value for each proposed sort criterion is determined based on the entity information contained in the catalog database.

The method according to the invention may comprise a display in a display window of the result of the sorting. This display window can be the one used to display the sorting criteria as well as the sorting values for each sort criterion.

Advantageously, the method according to the invention can comprise, during the various sorting steps, a display in a display window of the entities satisfying the sorting values selected from the sorting criteria.

The display of these entities can be performed either automatically or triggered by the user according to the result of the sorting.

The method according to the invention may further comprise updating the multidimensional database and the intermediate database at a predetermined frequency and / or controlled by an administrator. The update can be performed following, for example, a modification of the sorting criteria, the sorting values, the catalog database, etc.

The method according to the invention can advantageously be used for sorting product information, in an electronic catalog, offered for sale on a website. In this particular case, a category of products may be "camcorders", and the sorting criteria and the values may, for example, be as follows:

- price: <100, 100-200, 200-300,> 300 etc.

- brand: Sony, Nikkon, Toshiba, etc.

- recording media: "MiniDV", "DVD", "memory card", etc. According to another aspect of the invention there is provided a system implementing the method according to the invention. Such a system comprises a multidimensional database, means for displaying at least one sorting criterion and at least one sorting value per sorting criterion and means for accessing the multidimensional base as a function of the values. of selected sorting criteria. Thus, the system according to the invention may comprise an MDX request generator for accessing the multidimensional database. The database as well as the MDX query builder can be on a server. The system according to the invention may further comprise an intermediate database accessible from the multidimensional database and a catalog database accessible from the intermediate database.

The system according to the invention may further comprise a back office computer tool for managing criteria and sorting values as well as the various databases and their update process.

Finally, the system according to the invention comprises means for linking the various databases, in the form of an SQL query, a computer program, etc.

Other advantages and features of the invention will appear on examining the detailed description of a non-limiting embodiment, and the attached drawings in which:

FIG. 1 is a schematic representation of an exemplary system according to the invention; Figure 2 is a schematic representation of the use of an OLAP cube according to the invention;

FIG. 3 is a representation of a list of sorting criteria;

FIG. 4 is a representation of display windows allowing an administrator to activate a sorting criterion and the sorting values for this sorting criterion;

FIG. 5 is an example of displaying a multidimensional database according to the invention using a multidimensional database explorer; FIG. 6 is an example of display windows making it possible to perform a drilldown mode sorting, the result of which is determined according to the invention; and FIG. 7 is an example of display windows making it possible to carry out an "expert" sorting, the result of which is determined according to the invention;

The particular example of application that we will describe in the following concerns the sorting of product information offered for sale, in an electronic catalog on a website.

FIG. 1 gives a schematic representation of an exemplary system according to the invention. The system represented in FIG. 1 comprises a multidimensional database 11, the content of which is calculated from an intermediate database 12 constituting the data source for the multidimensional database 11. The intermediate database 12 is constructed at from data contained in a catalog database 13 corresponding to the electronic catalog containing the products that may be affected by a sort of information made by a user. The system further comprises a Back Office tool 14 allowing an administrator to configure the sorting criteria, the sorting values for each sorting criterion, and the construction of the various databases, their updating, as well as the general operation of the system.

The multidimensional database comprises a plurality of OLAP cubes 111, each corresponding to a category or a family of products such as, for example, microwaves, mp3 players, etc. Each sort criterion gives rise to an analysis axis corresponding to one dimension in each OLAP cube. Each sort value for a sort criterion gives rise to a member on the dimension corresponding to the sorting criterion. The result of sorting is the number of products corresponding to a sorting value of a sorting criterion. The OLAP database will therefore contain all the product numbers at the intersection of all possible combinations of all the values of all the sorting criteria.

The database is accessed by Multi Dimensional Expressions (MDX) queries. These queries are generated by a generator of MDX query 15 from the sort criteria and the values selected for each sort criterion. The MDX language is the language associated with the multidimensional databases used to query MS SQL Server. The MDX query generator dynamically interrogates the multidimensional database to render a new list of sorting criteria and sorting values available for these sorting criteria so as to perform a new sorting while avoiding having a sorting result. no.

The system further includes an SQL query generator 16 querying the catalog database to display the products corresponding to the sorting result.

To better assist a user, or user, in its process of choosing a product in the electronic catalog and to improve its experience in browsing the catalog, the method according to the invention uses all the information relating to products such as as the brand, the price, the size, the promo code, the color, etc. This information is called product characteristics. In the rest of the description we will call filter a sort value for a sort criterion.

We will now describe an example of implementation of the invention. In the following description "filter" corresponds to a sort value for a sort criterion.

Choice of sorting criteria and filters

According to the present invention, each catalog gives rise to a multidimensional database 11. Each product family gives rise to an OLAP Cube 111. Each sort criterion gives rise to an analysis axis or an OLAP dimension. The result of sorting is the number of products corresponding to a sorting value of a sorting criterion. The multidimensional database 11 therefore contains the set of product numbers at the intersection of all possible combinations of all the values of all the sorting criteria.

FIG. 2 gives a schematic representation of an OLAP cube 111. The dimension 21, 22, 23 each correspond to a selection criterion such as for example the color, the mark and the price. Each dimension 21, 22, 23, to a default member, respectively 211, 221, 231, which is the "all" or "AlIProducts" member corresponding to the total number of products in the product family associated with that cube. The dimensions 21-23 also have other members corresponding to sort values for the sorting criteria associated with dimensions 21-23. For example, dimension 21, which corresponds to the "Color" criterion, has two other members: o member 212 is associated with the sorting value "Red" and o member 213 is associated with the sorting value "Blue" D ' similarly, the dimensions 22 and 23, associated respectively with the sorting criteria "Mark" and "Price", have members 222, 223 and 232, 233, 234 which correspond to sorting values for the price and brand sorting criteria. :

- 222: Sony, and 223: Toshiba;

- 232: Less than 100 €, 233: from 100 to 200 €, and 234: more than 200 €.

So when a first sort is made with the following combination: o Brand = Sony; o Color = Red; and o Price = less than 100 € The result of sorting is given by the content 24 of the cube 111, that is to say 5. There are therefore 5 products satisfying the criteria and When another sorting value for a sorting criterion is added, for example Price = Less than 100 € and 100 to 200 €, then the result of the sorting is given by the sum of the contents 24 and 25 of the cube 111, that is to say 5 + 2 = 7. There are 7 products that satisfy all the sort values selected for the sort criteria.

The filters offered to the user are set in the Back Office tool 14. The filters can be set up according to several steps which are as follows:

Step 1: Activation of a characteristic present in the catalog or catalog database.

Step 2: Choice of data type: nvarchar, int, decimal, datetime or bit. Step 3: Activation of a characteristic in sorting criteria: the administrator decides to include the characteristic in the proposed criteria to the user:

- in a sorting mode "drilldown" or single-criteria, or - in a sorting mode "expert" or multi-criteria

Step 4: Setting the value ranges if the characteristic is numeric (decimal or int) or Datetime type Step 5: Scheduling of the sort criteria in drilldown mode and expert Step 6: Publication and generation of data.

Figure 3 shows a window 31 summarizing the characteristics that can be activated by the administrator in sorting criteria. Each characteristic is symbolized by a cylinder. For each characteristic, the administrator can access the contents of the catalog for a product family and view the values associated with this characteristic and the number of occurrences of each value. These values are presented in a table 32, make it easier for the administrator to choose the filters associated with this characteristic.

Then, using a window 41, shown in Figure 4, the administrator activates the feature, changes its label, and indicates whether it will be visible in different contexts:

- the product sheet,

- the comparison table,

- the list of the results of a search, in a sort in autocompletion, that is to say in a mode where the user can carry out an assisted text search. For example, as a simple text box, the user is prompted to search for a product. As and when it hits, the relevant answers are presented. - in simple sorting criterion in "drilldown" mode, and

- in advanced sorting criterion in expert mode. The administrator also performs the typing, using a window

42 is still represented in FIG. 4. If it is a numerical criterion, the user can also specify the ranges of values to be presented to Internet users during sorting operations. In the case of the example presented in FIG. 4, the values and ranges of values specified in the window 42 generate the filters indicated in the window 43.

The order of appearance of the sorting criteria can also be administered in the Back Office tool 14.

To populate the multidimensional database 11, the intermediate database 12 is generated from the catalog database 13. This intermediate database 12 is in fact the data source for the multidimensional database 11.

Creating the Intermediate Base 12

In the intermediate database 12, several modes of presentation of the filters are possible. The values presented can be of the following form: - Less than XXX

- Between X and Y Superior to YYY

- Equal to Z

Keywords are used to represent these operators: - Inf

Between

- Sup

- Equal

The storage of the values in the intermediate data base 12 is done in this form with the character "pipe" as separator: Operator | val l | val2 Example: Equal | Sony | Between I 100 | 150 At the time of publishing a family (after configuration), this data model is regenerated to take into account the latest values for each criterion and for each product, to integrate the new sorting characteristics, to adapt ranges of values presented. In the present example, the intermediate database 12 is constructed according to a single table scheme by family containing: the product identifier a column by characteristic parameterized as a filter criterion. The intermediate database 12 may also be constructed in a "star" pattern for better performance when feeding the multidimensional database 11, i.e., one table per feature. But the performance imperatives are at the time of queries during the sorting performed by the users and not when feeding the multidimensional database 11.

When publishing data, a [C_CreateFamiCubeTable] procedure is invoked for each product family. This procedure performs:

the creation of a table with the set of fields for each product family, and the supply of the intermediate database 12 from the catalog database 13.

The creation of the table is based on all the characteristics that can be used in filter criteria (drilldown or expert mode). The columns corresponding to the set of characteristics selected by the administrator as sorting criterion are created for each product family of the intermediate database

12. With the exception of the product identifier, all fields are typed "string". The model of a table by product family is therefore the following:

Prodld: product identifier

ProdK [Characteristic Identifier]: Calculated product value for the characteristic. ProdKXXX .... Each table has an identifier constructed as follows: X [family identifier].

The intermediate table is fed by using a code in Assembly.NET (compiled binary code) written in C # "CaraPlagBuilder" referenced in the SQL Server engine as a GetPlag function:

CREATE FUNCTION [dbo] [GetPlag] @CaraValue [nvarchar] @CaraPlag [int] @ CaraMιnιPlag [nvarchar] JOOO @CaraType [nvarchar], 50 ^* RETURNS [nvarchar] 1000 WITH EXECUTE AS CALLER EXTERNAL NAME [CaraPlagBuilder] [CaraPlagBuilder] .PlagBuilder] [GetPlag]

This guarantees optimal performance for the treatments and calculations performed. The parameters are as follows: - @CaraValue: Current value for the characteristic for the product

- @CaraPlag: number of desired value ranges (optional)

- @CaraMiniPlag: value range terminals

- @CaraType: parameterized data type To guarantee the internationalization of a sorting engine implementing the system according to the invention, and to ensure the independence of the language, the values are first "cleaned" of all the special characters possible. Thus the following characters: Â à à a a a â é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é é?

AAAAAAaaaaaaOOOOOOooooooEEEEeeeeiiiIiiiMUUUUuuuuyNπCc If the value of the characteristic is empty string or NULL or 0, the return value will be "equal | | ". If there are no defined ranges, the returned value will be "equal | va / eur | ". If ranges have been specified in parameter, the various specified ranges are searched to find to which the product belongs. If the product is below the first range value, the return value will be "inf | .Z ^era value \ ". If the product is above the last range value, the return value will be "sup I last value \". If the product is between 2 terminals, the return value will be "between | bornel | terminal2 ". Each product belongs to only one dimension member and therefore is located in only one range of characteristic values. The mass call of this function thus makes it possible to feed the source table, in the intermediate database, from which the OLAP cubes of the multidimensional database will be fed. The table below gives a example source table in the intermediate database 12 for the "camcorder" product family.

equal | DVD-R 3 ", 3" DVD- equal I AC adapter

92502 inf | 3 | RW, 3 "DVD + RW, 3" external I DVD + R DL, 3 "DVD + R | equal I AC adapter

DVD + R DL | equal I DVD-R 3 ", 3" DVD- equal I AC adapter

92521 between | 4 | 6

RW I external I equal I 13 "equal power adapter DVD-RAM, DVD-R

92522 between | 4 | 6 + battery charger - 3 ", 3" external DVD-RWI I equal I Equivalent 13 "power adapter DVD-RAM, DVD-R

92528 inf | 3 | + battery charger - 3 ", 3" external DVD-RWI I equal I AC adapter

92540 between | 6 | 8 equal | Mini DV | external I

From the generated data, the multidimensional database 11 can be powered. Each product family having its own characteristics parameterized as a filter criterion, an independent cube 111 per family is necessary. A module named "CCubeGenerator" allows to build the structure of each cube 111. It is called to each family publication to refresh:

the list of characteristics parameterized in sorting criterion, therefore the list of dimensions of the corresponding cube; - the values presented (the ranges or raw values) for each product, therefore the membership of each product to the members of each dimension; the numbers of products located at the intersection of each member of each dimension and their aggregates.

The Microsoft object model. AnalysisServices provides programmatic access to the structure of SQL multidimensional databases

Server and thus to create or modify databases in an automated way.

Thus the different steps carried out at this stage are as follows:

Step 1: Creating the multidimensional database 11:

If the multidimensional database for the catalog does not exist, it is created.

Step 2: Remove a cube 111

If the current family already has a cube 111, delete it. Step 3: Creating the data source

Refreshing the list of available tables in the intermediate database for integration into cube 111. Added table X [family identifier] to the data diagram.

Step 4: Creating the new dimensions [AII Products] 1. Definition of the default member, the structure and the storage mode: dim.Type = Λ ^v x ^ ^N - sx .Regular; dim.AttributeAIIMemberName = "AH Products" The default member is the set of products for a sort criterion.

2. Creating the root attribute based on the identifier:

3. Retrieval of the list of characteristics in sorting criteria. For each of the criteria, create a dimension whose members will be the possible distinct values. Connection to the Source field in the intermediate database.

4. Create the default dimension hierarchy based on the "Prodld" attribute.

Step 5: Creating / Editing a cube 111 Added the COUNT metric that corresponds to the number of products that meet the user's filter criteria. Save the new configuration of Cube 111. Step 6: Refresh Data This is a step of incremental or total update of multidimensional database 11.

Fig. 5 is an example of a configuration of a multidimensional database 11 obtained with a database explorer.

The essential problem here is the performance, because:

the number of requests to be made in order to permanently obtain the number of products corresponding to each criterion value

- the number of Internet users potentially present at the same time - the Internet technologies themselves: a simple and stateless protocol (http), the multitude of browsers clients (Firefox, Internet Explorer, Opera ..) and their versions, the systems operating system (Linux, Windows ...) and embedded client development technologies (javascript) Since the number of criteria is greater than 5 and can reach 25 or even 30 criteria, a precalculation / client cache solution and / or server would be tedious, very complex and particularly cumbersome to implement.

The system according to the invention which has just been described has the following advantages:

- Speed of calculation,

- Aggregates and composition facilitated,

- Assured increase in charge,

- Ease of use The multidimensional database 11 is a database made with OLAP "Analysis Services" modules of the Microsoft SQL Server 2005 engine.

We will now describe with reference to Figures 6 and 7 two examples of sorting with the system according to the invention. Figure 6 presents three display windows, 61-63, presenting the different steps of an example of sorting in drilldown or single-criteria mode. The example of sorting relates to "microwaves. The window 61 corresponds to the starting point of the sorting. Four first sorting criteria are proposed: brand, price, type and volume. The filters, or sorting value, proposed for these sorting criteria are given below for each sorting criterion:

- Brand: Ariston, Moulinex, Panasonic and Russell Hobbs;

- Price: less than 100 Euros, from 10 to 200 Euros, 200 to 300 Euros and 300 to 400 Euros; - Type: Microwave and Grill, Microwave convection and convection, Microwave convection and Convection, Microwave; and

- Volume: 10 to 20 liters, 20 to 30 liters, 30 to 40 liters, and more than 40 liters; and For each filter, the corresponding product number is displayed in parentheses next to this filter. The total number of microwaves is displayed at the top left of window 61.

The user clicks on the filter "from 20 to 30 liters". The total number of microwaves satisfying the "Volume from 20 to 30 liters" filter is automatically displayed at the top left of window 62 as well as the "Volume from 20 to 30 liters" filter. A new sorting criterion, or filter criterion is then proposed on the window 62: "the power". All combinations of possible filters are then recalculated and the number of products corresponding to these combinations. Only the available filters, that is, resulting in a non-zero result, are displayed with the number of microwaves corresponding to these combinations. The filters available for the "Power" sorting criterion and the number of microwaves satisfying these filters are:

- from 500 to 1000 watts (11), and - from 1000 to 1500 watts (1).

The user clicks on the filter "from 100 to 200 €". The combinations are recalculated and displayed on the window 63. The combination of filters chosen by the user, namely "Volume 100 to 200" and price 100 to 200 € "and the number of microwaves satisfying this combination Filters are displayed at the top left of window 63. Note that the Ariston brand is no longer offered. This comes from the fact that no Ariston brand microwaves correspond to the combination of filters "Volume from 100 to 200" and Price from 100 to 200 € ». In addition, a new criterion appears: "rotating heat". In addition, a new filter for the "Brand" criterion appears: "Samsung". And so on.

In the example just described, the sorting criteria as well as the filters for these sorting criteria are presented in "drilldown" mode, that is to say in the form of an updated tree with reloading. of the display window. In this drilldown mode only one filter can be selected at each step of the sorting.

We will now describe a second example with reference to FIG. 7, which presents three display windows 71-73 of three stages of an expert or multi-criteria sorting on a category of products corresponding to mp3 players. In this second example, the user can select multiple filters at a time. For each of the sorting criteria, the filters are presented in a drop-down menu that the user can control. Whatever the criterion, the invention makes it possible to display the various available filters as well as the number of products corresponding to each filter. The total number of products satisfying the selected filter combination is also calculated and displayed instantaneously at the top left of the windows 71 to 73. On the window 71, the user selects the filter "Indifferent", the total number of mp3 players satisfying this The filter is displayed in real time: 56. On the window 72, the user has decreased the drop-down menu corresponding to the "brand" criterion and has rolled out the drop-down menu corresponding to the "Price" criterion. The number of mp3 players satisfying each of the filters for the price criterion is calculated and displayed in parentheses. The user selects the filters "50 to 100 €" and "100 to 150 €". The total number of mp3 players corresponding is displayed in real time. The user can return directly to one of the filters he has selected during the different stages of sorting without necessarily respecting the order of the steps. On the display window 73, the user has deselected the price filter "from 50 to 100 €". It has reduced the drop-down menu corresponding to the "price" criterion, and has new scroll down the drop-down menu corresponding to the criterion "mark". Note that the price filter "from 100 to 150 €" is displayed at the "price" criterion. In this "expert" mode, the user can therefore perform a quick search, while remaining assured that at least one product corresponds to the combination of filters he has selected. It can easily return to some of the filter criteria very easily, quickly, simply by deselecting them, and without having to rebuild the filter combination from the beginning. It has a button 74 "Show results" allowing him to display mp3 players satisfying the combination of filters he has selected.

The examples presented above show that the invention makes it possible to:

- facilitate the process of sorting and choosing a product by offering a high number of criteria for both specialists and novice Internet users,

- permanently and dynamically display the number of products remaining following the filters applied by the user.

- Ensure the availability of at least one product regardless of the sorting values of the sorting criteria chosen by the user. - Perform sorting in very good response times, for example less than one second, in all circumstances and despite the number of information to be processed.

The invention is not limited to the examples which have just been described and can be applied to the sorting of all entities that can be grouped into at least one category or family according to the information relating to these entities.

Claims

A method of sorting information relating to a plurality of entities, according to at least one sorting value associated with at least one sorting criterion: said sorting criterion being selected from among a plurality of sorting criteria each relating to a characteristic of said entities, and having one or more sorting criterion values, and said sorting value being selected from one or more sorting values of said sorting criterion; characterized in that

each sorting criterion is associated with a dimension of a multidimensional database (11), said database being constructed from data relating to said entities, each of the sorting values of a sorting criterion being associated with a member of the dimension associated with said sorting criterion, and

the result of said sorting is determined according to the content of the multidimensional database (11) having as coordinates the member associated with the sorting value on the dimension corresponding to said sorting criterion.

2. Method according to claim 1, characterized in that it comprises for a previous sorting performed according to at least one previous sorting value, a determination of at least one sorting value, said available, for at least one other criterion of sorting, which can be combined with said previous sorting value, to perform a new sort giving a non-zero result.

3. Method according to claim 2, characterized in that it further comprises a display of the result of a new sort for at least one available sorting value.

4. Method according to any one of the preceding claims, characterized in that the result of a sorting, carried out for several sorting values selected for one or more sort criteria, is determined according to the content of the multidimensional base (11). ) whose coordinates are the members associated with each of said sorting values on all dimensions associated with each of said sorting criteria.

5. Method according to any one of the preceding claims, characterized in that the sorting result comprises the number of entities satisfying the sorting values for the sorting criteria.

6. Method according to any one of the preceding claims, characterized in that the multidimensional database (11) is accessed by a request of the Multi Dimensional Expression (MDX) type relating to the combination of selected criteria of sorting criteria.

7. Method according to any one of the preceding claims, characterized in that the multidimensional database (11) is constructed from a database (12), called intermediate, comprising for each entity an identifier and a value sorting for each sort criterion.

Method according to claim 7, characterized in that the intermediate database (12) is constructed from a database

(13), called catalog, using SQL type queries comprising for each product its identifier and the sorting value corresponding to all sorting criteria.

9. Method according to any one of the preceding claims, characterized in that it further comprises a display in a display window (51-53, 61-63) of the sorting result.

10. Method according to any one of the preceding claims, characterized in that it further comprises a display in a display window entities satisfying the sorting values selected sorting criteria.

11. Method according to any one of the claims, characterized in that it further comprises a modification, by an administrator of at least one sorting criterion or at least one sorting value of a sorting criterion.

12. Method according to any one of the preceding claims, characterized in that it further comprises an update of the multidimensional base (11) at a predetermined frequency and / or controlled by an administrator.

13. Use of the method according to any one of the preceding claims for sorting products, in an electronic catalog, offered for sale on an Internet site.

A system implementing the method of any one of the preceding claims.