Importing data using metadata

[0001] The described technology relates to importing data into a database.

[0002] The World Wide Web is conducive to conducting electronic commerce (“e-commerce”). E-commerce generally refers to commercial transactions that are at least partially conducted using the World Wide Web. For example, numerous web sites are available through which a user using a web browser can purchase items such as books, groceries, and software. A user of these web sites can browse through an electronic catalog of available items to select the items to be purchased. To purchase the items, a user typically adds the items to an electronic shopping cart and then electronically pays for the items in the shopping cart. The purchased items can then be delivered to the user via conventional distribution channels (e.g., an overnight courier) or via electronic delivery when, for example, software is being purchased. Many web sites are also informational rather than commercial in nature. For example, many standards organizations and governmental organizations have web sites with a primary purpose of distributing information. Also, some web sites (e.g., a search engine) provide information and derive revenue from advertisements that are displayed.

[0003] The success of any web-based business depends in large part on the number of users who visit the business's web site and that number depends in large part on the usefulness and ease-of-use of the web site. Web sites typically collect extensive information on how their users use the site's web pages. This information may include a complete history of each HTTP request received by and each HTTP response sent by the web site. The web site may store this information in a navigation file, also referred to as a log file or click stream file. By analyzing this navigation information, a web site operator may be able to identify trends in the access of the web pages and modify the web site to make it easier to navigate and more useful. Because the information is presented as a series of events that are not sorted in a useful way, many software tools are available to assist in this analysis. A web site operator would typically purchase such a tool and install it on one of the computers of the web site. There are several drawbacks with the use of such an approach of analyzing navigation information. First, the analysis often is given a low priority because the programmers are typically busy with the high priority task of maintaining the web site. Second, the tools that are available provide little more than standard reports relating to low-level navigation through a web site. Such reports are not very useful in helping a web site operator to visualize and discover high-level access trends. Recognition of these high-level access trends can help a web site operator to design the web site. Third, web sites are typically resource intensive, that is they use a lot of computing resources and may not have available resources to effectively analyze the navigation information.

[0004] It would also be useful to analyze the execution of computer programs other than web server programs. In particular, many types of computer programs generate events that are logged by the computer programs themselves or by other programs that receive the events. If a computer program does not generate explicit events, another program may be able to monitor the execution and generate events on behalf of that computer program. Regardless of how event data is collected, it may be important to analyze that data. For example, the developer of an operating system may want to track and analyze how the operating system is used so that the developer can focus resources on problems that are detected, optimize services that are frequently accessed, and so on. The operating system may generate a log file that contains entries for various types of events (e.g., invocation of a certain system call).

[0005] To overcome some of the drawbacks mentioned above, a hosted service may be provided that collects event information from various web sites, stores the information at a central location, and provides a web site through which web site operators can access and analyze the information collected from their web site. Because the format and organization of the data collected from different web sites may be very different, the hosted service provider may need to expend considerable effort to develop conversion programs to convert the collected information into a format that is compatible with the data warehouse of the hosted service. It would be desirable to reduce the effort needed to convert the collected information.

[0027] A method and system for importing data into a data store in accordance with metadata is provided. In one embodiment, the import system provides metadata that specifies how the import data for various types of import sources is to be imported into the data store. The import sources may be categorized according to the type of data provided by the import sources. For example, various web sites may provide similar types of data (e.g., click stream data) that may be categorized as the same type of import source. Alternatively, each company or organization may be categorized as its own type of import source. When the import system receives the import data from the import source, it identifies the type of import source and retrieves the metadata defined for that type of import source. The import system then imports the received import data into the data store in accordance with the retrieved metadata. The retrieved metadata may define a schema for the data store and define how the import data is to be stored in the data store. The schema may define the various tables for the data store. The schema may include for each table an indication of the properties (or columns) for that table. The import system creates the tables in accordance with the schema and then imports the import data into the created tables. The metadata may indicate, for each property of a table, how the value for that property is to be calculated from a field of the import data. In this way, metadata for the same type of import source can be shared by import sources of that type and the metadata can specify the schema of the data store. The use of metadata reduces the effort needed to import data into the data store.

[0028] FIG. 1 is a block diagram illustrating the organization of the import system in one embodiment. In this embodiment, the import system includes log data 101, parser 102, event data 103, miscellaneous data 104, import engine 105, import metadata 106, data warehouse (“DW”) 107 (i.e., a data store), aggregation engine 108, aggregation metadata 109, and aggregated data 110. The log data may represent a click stream file generated by a web server. The parser processes the log data and generates various types of event data. For example, event data may include the identification of the click stream associated with each visit of a user to the web server. U.S. patent application Ser. No. 09/613,847, entitled “Method and System for Parsing Navigation Information,” which is hereby incorporated by reference, describes a parser. The miscellaneous data may include information on products available to be purchased by users via the web server and the identification of registered users of the web server. The import metadata describes how the event and miscellaneous data, referred to collectively as “import data,” are to be imported by the import engine and stored in the data warehouse. The import metadata not only defines how the import data is to be transformed, but also defines the structure of the tables of the data warehouse in which the import data is to be stored. The aggregation metadata defines how the data in the data warehouse is to be aggregated. The aggregation engine inputs the aggregation metadata, aggregates the data of the data warehouse in accordance with the aggregation metadata, and stores the aggregated data. One skilled in the art would appreciate that the aggregated data may be stored in the data warehouse itself. The import system may execute on a computer system that includes main memory, a central processing unit, input devices (e.g., keyboard and pointing devices), output devices (e.g., display devices), and a hard drive. The main memory and hard drive are computer-readable media that may contain instructions that implement the import system. The import data may receive data from various communication channels that includes the Internet, a wide area network, a point-to-point dial-up connection, or a local area network.

[0029] FIG. 2 is a block diagram illustrating the structure of the import metadata. The import metadata is stored in a meta_—Dimension table 201 and a meta_—DimProperty table 202. The meta_—Dimension table contains an entry for each table of the data warehouse into which data is to be imported. The meta_—DimProperty table contains an entry for each column or property of each table specified by the meta_—Dimension table. Each entry of the meta_—DimProperty table contains import data 203 and schema data 204. Tables 1–4 describe the contents of the meta_—Dimension table and the meta_—DimProperty table.

[0030] The meta_—Dimension table contains an entry for each table of the data warehouse into which data is to be imported.

[0031]

meta—Dimension
AttributeName	Description
DimName	Name of the table of the data warehouse
IsAddVisitInformation	Indicates that visit identification information is to
	be added to this table (“this” table refers to the
	table of the data warehouse described by this
	entry in the meta—Dimension table)
UpdateParent	Name of the parent table of this table to be
	updated
Expression	Code that indicates how the parent table is to be
	updated
UseWorktables	Indicates whether the import data for this table
	should be stored in a temporary table before being
	moved to this table in the data warehouse
DeleteAndImport	Indicates that all rows of this table in the data
	warehouse with the same identifier as a row to be
	imported are to be deleted before the row is
	imported
CreateLineNos	Indicates that line number data is to be added the
	rows of this table

[0032] The meta_—DimProperty table contains an entry for each column of each table in the data warehouse. The combination of the DimName and PropertyName attributes uniquely identifies the entries of this table.

[0033]

meta—DimProperty
AttributeName	Description
DimName	Name of this table in the data warehouse
	(corresponds to the DimName attribute of the
	meta—Dimension table)
PropertyName	Name of this column of this table in the data
	warehouse
PropertyDisplayName	Display name of this column
ImportData	see meta—DimProperty (import data)
SchemaData	see meta—DimProperty (schema data)

[0034] The import data attributes of the meta_—DimProperty table describes the data to be imported.

[0035]

meta—DimProperty (import data)
AttributeName	Description
input—PropertyName	Name of column of import data corresponding
	to this column
input—SqlDataType—Def	Type of data in this column of the import data
	(e.g., integer)
input—ColumnNumber	Column number of this column in the import
	data
SqlType	Specifies the SQL type of this column
SqlType—Def	Specifies the SQL type definition of this
	column
SqlType—Length	Specifies the SQL type length of this column
SqlType—Precision	Specifies the SQL type precision of this column
SqlType—Scale	Specifies the SQL type scale of this column
SqlType—AllowNulls	Specifies whether nulls are allowed for the
	SQL type of this column
SqlType—DefaultValues	Specifies default values for the SQL type of
	this column
TransformationString	Code to transform the import data for this
	column

[0036] The schema data attributes of the meta_—DimProperty table describe how to store the import data in this column in this table.

[0037]

meta—DimProperty (schema data)
AttributeName	Description
IsAddedToSchema	Indicates whether this column has already be added
	to this table in the data warehouse
IsLookup	Indicates whether a lookup table should be
	created for this column
IsDerived	Indicates whether the TransformationString
	attribute should be used to transform the import
	data for this column
IsStatic	Indicates whether the history of this column
	should be tracked
IsDaily	Indicates whether this column is to be aggregated
	on a daily basis
IsWeekly	Indicates whether this column is to be aggregated
	on a weekly basis
IsMonthly	Indicates whether this column is to be aggregated
	on a monthly basis
IsMultiValued	Indicates whether this column contains an array
IsAggregated	Indicates whether this column is to be aggregated
IsHash	Indicates whether this column should be included
	in a hash value to determine whether a row of this
	table has changed
IsIdentifiable	Indicates whether this column is part of the
	unique identifier for this table
ChangeRateBitMap	Combination of the IsDaily, IsWeekly, and
	IsMonthly attributes
Reldimension	Indicates the table of the data warehouse that
	contains dimension data for this column (i.e., the
	related dimension table)
IsContinuous	Indicates whether the data of this column should
	be placed in buckets
Timeoffset	Seconds to add to the import data for this column

[0038] FIGS. 3–13 are flow diagrams illustrating the processing of the import system in one embodiment. FIG. 3 is a flow diagram of the processing of the import engine in one embodiment. The import engine is passed the import data and the name of the table of the data warehouse into which the data is to be imported. The import engine may be invoked for each set of import data that is to be imported. The processing in each of the blocks of this flow diagram is further described in another flow diagram. In block 301, the engine creates a temporary table to store the import data before it is transferred to the data warehouse. In one embodiment, the engine may store the data directly in the table of the data warehouse based on the setting of the UseWorktables attribute of the meta_—Dimension table. In block 302, the engine copies the import data to the temporary table. In block 303, the engine calculates the hash value for each of the rows of the temporary table. The hash values are used to help identify whether one row contains the same data as another row. For example, if a row to be imported contains the same data as a row currently in the table of the data warehouse. In block 304, the engine checks for duplicate data in the import data and removes the duplicate data. In block 305, the engine creates lookup tables for the columns of the table of the data warehouse. In block 306, the engine creates buckets for the columns whose IsContinuous attribute is set. In block 307, the engine resolves the dimensions in the temporary table. For example, if one of the columns of the table of the data warehouse holds a reference to a product table, then the rows of the temporary table are set to include a reference to the product table. In block 308, the engine updates the data warehouse based on the information in the temporary table. In block 309, the engine updates the visit mappings. That is, if the IsAddVisitInformation attribute of the meta_—Dimension table is set, then the engine calculates and adds the corresponding visit identifier for each row added to the table of the data warehouse. In block 310, if the UpdateParent attribute of the meta_—Dimension table is set, then the engine updates the parent table of this table in the data warehouse. The engine then completes its processing.

[0039] FIG. 4 is a flow diagram of the processing of the create temporary table function in one embodiment. This function is passed the name of the table and creates a temporary table with a column for each property of that table. In blocks 401–403, the function loops selecting each of the properties of the table and recording the column name for the temporary table. In block 401, the function selects the next property name of the meta_—DimProperty table. In decision block 402, if all the property names have already been selected for the passed table, then the function continues at block 404, else the function continues at block 403. In block 403, the function adds a column for the selected property name to the temporary table and then loops to block 401 to select the next property name. In block 404, the function requests the creation of the temporary table with all of the columns identified in block 403. The function then returns. This function may also add other columns corresponding to the columns of the import data. For example, if the IsContinuous attribute is set for a property, then the function may add an additional column to the temporary table for the corresponding import value.

[0040] FIG. 5 is a flow diagram of the processing of the copy import data to temporary data function in one embodiment. This function copies each row of the import data to the temporary table and generates the derived values as appropriate. In blocks 501–509, the function loops, selecting and processing each row of the import data. In block 501, the function selects the next row of the import data. In decision block 502, if all the rows of the import data have already been selected, then the function returns, else the function continues at block 503. In blocks 503–508, the function loops, selecting each property of the table of the data warehouse and generating a value to store in the temporary table. In block 503, the function selects the next entry for the meta_—DimProperty table. In decision block 504, if all the entries of the meta_—DimProperty table have already been selected, then the function continues at block 509, else the function continues at block 505. In decision block 505, if the IsDerived flag is set for the selected entry, then the function continues at block 506, else the function continues at block 507. In block 506, the function generates the derived value using the TransformString attribute of the selected entry. In block 507, the function retrieves the value indicated by the ColumnNumber attribute from the selected row of import data. In block 508, the function adds the value to a row to be added to the temporary table and then loops to block 503 to select the next entry for the meta_—DimProperty table. In block 509, the function adds the row to the temporary table that contains the values added in block 508. The function then loops to block 501 to select the next row of the import data. One skilled in the art will appreciate that the temporary table may contain a separate column for the value of the import data when the IsDerived attribute is set.

[0041] FIG. 6 is a flow diagram of the processing of the calculate hash values function in one embodiment. The hash values are generated from the columns of the temporary table whose IsHash attribute is set. In block 601, the function selects the next row of the temporary table. In decision block 602, if all the rows of the temporary table have already been selected, then the function returns, else the function continues at block 603. In block 603, the function creates the hash value for the selected row by combining value of the columns whose IsHash attribute is set in the corresponding entries of the meta_—DimProperty table. In block 604, the function stores the created hash value in the selected row of the temporary table. The function then loops to block 601 to select the next row of the temporary table. One skilled in the art will appreciate that the import engine would add a hash column to any table of the data warehouse with at least one property whose IsHash attribute is set.

[0042] FIG. 7 is a flow diagram of the processing of the check for duplicate import data function. This function checks for rows of the import data that contain the same information and deletes all but the last row. The rows are considered duplicates if the values in their columns with the IsIdentifiable attribute set are the same. That is, the combination of the values from the properties whose IsIdentifiable attribute is set is the key for the row, which may be unique. In block 701, the function selects the next row of the temporary table. In decision block 702, if all the rows of the temporary table have already been selected, then the function returns, else the function continues at block 703. In block 703, the function locates the duplicate rows. In decision block 704, if duplicate rows are located, then the function continues at block 705, else the function loops to block 701 to select the next row of the temporary table. In block 705, the function deletes all but the last duplicate row from the temporary table and then loops to block 701 to select the next row of the temporary table.

[0043] FIG. 8 is a flow diagram of the processing of the create lookup tables function. This function loops, selecting each entry of the meta_—DimProperty table that has its IsLookup attribute set and creating a lookup table for each such entry. The lookup tables may have names derived from the property name (e.g., “lookup_—propertyname”). Lookup tables are used to represent import data that may be more compactly represented in an index to a table that contains the import data. For example, a property that contains the URL of accessed web pages (e.g., “www.acme.com/homepage”) would be a candidate property for a lookup table. If the web site provides 16 different web pages, then the index into the lookup table can be represented by four bits, rather than, for example, 32 bytes of the URL. In block 801, the function selects the next entry of the meta_—DimProperty table. In decision block 802, if all the entries of the meta_—DimProperty table have already been selected, then the function returns, else the function continues at block 803. In block 803, the function creates a lookup table for the selected entry. In blocks 804–809, the function loops, selecting each row of the temporary table and adding the appropriate entries to the lookup table. In block 804, the function selects the next row of the temporary table. In decision block 805, if all the rows of the temporary table have already been selected, then the function loops to block 801 to select the next entry of the meta_—DimProperty table, else the function continues at block 806. In block 806, the function retrieves the entry from the lookup table for the selected row of the temporary table. In decision block 807, if an entry was retrieved, then the function continues at block 809, else the function continues at block 808. In block 808, the function adds an entry to the lookup table corresponding to the value of the selected property for the selected row. In block 809, the function adds the index of the retrieved or added entry to the selected row and then loops to block 804 to select the next row of the temporary table. One skilled in the art will appreciate that the temporary table may include a column for holding the import value associated with each property whose IsLookup attribute is set.

[0044] FIG. 9 is a flow diagram of the processing of the create buckets function in one embodiment. This function creates bucket values for each column of the target table whose IsContinuous attribute is set. The bucket tables may be predefined tables with a name that is derived from the name of the selected property (e.g., “bucket_—propertyname”). Each entry of a bucket table may specify the range associated with that bucket. For example, if the property corresponds to the salary of employees, then the first bucket may have a range that is less than $25,000, the second bucket may have a range from $25,000 to $50,000 and so on. The function would assign an index of 2 when the salary is $35,000. In blocks 901–906, the function loops, selecting each entry of the meta_—DimProperty table whose IsContinuous attribute is set and identifying its bucket index. In block 901, the function selects the next entry of the meta_—DimProperty table whose IsContinuous attribute is set. In decision block 902, if all such entries of the meta_—DimProperty table have already been selected, then the function returns, else the function continues at block 903. In blocks 903–906, the function loops, selecting each row of the temporary table and updating its bucket index for the selected entry of the meta_—DimProperty table. In block 903, the function selects the next row of the temporary table. In decision block 904, if all the rows of the temporary table have already been selected, then the function loops to block 901 to select the next entry of the meta_—DimProperty table, else the function continues at block 905. In block 905, the function looks up the bucket index for the value of the selected row. In block 906, the function adds the bucket index to the selected row and loops to block 903 to select the next row of the temporary table. One skilled in the art would appreciate that the temporary table may contain an additional column for holding the import data corresponding to each property whose IsContinuous attribute is set.

[0045] FIG. 10 is a flow diagram of the processing of the resolve dimensions function in one embodiment. The resolve dimensions function updates the columns of the table of the data warehouse that correspond to a property whose RelDimension attribute indicates a dimension table. The function loops, selecting each entry of the meta_—DimProperty table. In block 1001, the function selects the next entry of the meta_—DimProperty table. In decision block 1002, if all the entries have already been selected, then the function returns, else the function continues at block 1003. In decision block 1003, if the RelDimension attribute of the selected entry names a dimension table, then the function continues at block 1004, else the function loops to block 1001 to select the next entry of the meta_—DimProperty table. In blocks 1004–1009, the function loops retrieving each row of the temporary table and resolving the values for the selected property. In block 1004, the function retrieves the next row of the temporary table. In decision block 1005, if all the rows of the temporary table have already been retrieved, then the function loops to block 1001 to select the next entry of the meta_—DimProperty table, else the function continues at block 1006. In block 1006, the function retrieves the entry from the related dimension table as indicated by the Reldimension attribute. In decision block 1007, if an entry was retrieved, then the function continues at block 1009, else the function continues at block 1008. In block 1008, the function adds an entry to the related dimension table. In block 1009, the function updates the column corresponding to the selected property in the selected row and then loops to block 1004 to select the next row of the temporary table.

[0046] FIG. 11 is a flow diagram of the update data warehouse function in one embodiment. This function copies the data of the temporary table to the table of the data warehouse. In block 1101, the function selects the next row of the temporary table. In decision block 1102, if all the rows of the temporary table have already been selected, then the function returns, else the function continues at block 1103. In decision block 1103, if the DeleteAndlmport attribute of the meta_—Dimension table entry is set, then the function continues at block 1104, else the function continues at block 1106. In block 1104, the function retrieves all rows of the data warehouse table with the same key (generated from properties whose IsIdentifiable attribute is set) as the selected row of the temporary table. In block 1105, the function deletes all the retrieved rows of the data warehouse table. In block 1106, the function adds the selected row of the temporary table to the data warehouse table. The function then loops to block 1101 to select the next row of the temporary table.

[0047] FIG. 12 is a flow diagram of the processing of the update visit mapping function in one embodiment. This function updates the visit reference of the table in the data warehouse if the IsAddVisitInformation attribute of the meta_—Dimension table is set. The import engine adds a visit reference column to each table whose IsAddVisitInformation attribute is set. In decision block 1201, if the IsAddVisitInformation attribute of the meta_—Dimension table is set, then the function continues at block 1202, else the function returns. In blocks 1202–1205, the function loops selecting each row of the temporary table and updating its visit reference. In block 1202, the function selects the next row of the temporary table. In decision block 1203, if all the rows of the temporary table have already been selected, then the function returns, else the function continues at block 1204. In block 1204, the function identifies the related visit reference. The function may identify related visits based on timing information stored in the visit table of the data warehouse. In block 1205, the function updates the row of the data warehouse table corresponding to the selected row of the temporary table with the visit reference and then loops to block 1202 to select the next row of the temporary table.

[0048] FIG. 13 is a flow diagram of the processing of the update parent function in one embodiment. This function updates the parent table in the data warehouse if the UpdateParent attribute contains the name of the parent table. The function updates the parent table by executing the code of the Expression attribute on each row of the temporary table. For example, the rows of the temporary table may correspond to line items of a purchase order and the parent table may contain a row for each purchase order. The expression may add the price for each line item to a total price column of the row for the purchase order in the parent table. In decision block 1301, if the UpdateParent attribute in the meta_—Dimension table contains the name of parent table, then the function continues at block 1302, else the function returns. In block 1302, the function selects the next row of the temporary table. In decision block 1303, if all the rows of the temporary table have already been selected, then the function returns, else the function continues at block 1304. In block 1304, the function executes the expression on the selected row, which updates the corresponding parent entry in the data warehouse. The expression is specified by the Expression attribute. The function then loops to block 1302 to select the next row of the temporary table.

[0049] FIG. 14 is a block diagram illustrating the generation of an aggregation and report table. The aggregation table 1401 includes an aggregating property in the first column and the various measure columns. The aggregation engine inputs of the aggregation metadata that defines the aggregating property and the measure columns and creates the aggregation table. The report table 1402 similarly includes an aggregating property in the first column and various measure columns as defined by the aggregation metadata.

[0050] FIG. 15 is a block diagram illustrating the organization of the aggregation metadata in one embodiment. The aggregation metadata includes a meta_—Aggregation table 1501, a meta_—AggregationFact table 1502, a meta_—AggregationMeasure table 1503, a meta_—FactMeasure table 1504, a meta_—AggregationProperty table 1505, a meta_—ReportMeasure table 1506, and a meta_—Fact table 1507. The meta_—Aggregation table contains an entry for each aggregation that is defined. The meta_—AggregationFact table contains for each aggregation an entry for each fact table that is used to generate that aggregation. For example, a single aggregation may be generated from the data of five different fact tables in which case the meta_—AggregationFact table would have five entries for that aggregation. The meta_—AggregationMeasure table contains for each entry of the meta_—AggregationFact table an entry for each measure of the aggregation that is generated from that fact table. The meta_—FactMeasure table has for each fact table an entry for each measure that is generated from the fact table and specifies how that measure is to be generated. The meta_—AggregationProperty table specifies the aggregating property for an aggregation. The meta_—ReportMeasure table specifies the measure names to be included in the report for the aggregation. The meta_—Fact table contains an entry for each fact table that is used in an aggregation. The following tables describe the contents of these tables.

[0051]

meta—Aggregation Table
	AttributeName	Description
	AggregationName	Name of this aggregation

[0052]

meta—AggregationFact Table
AttributeName	Description
AggregationName	Name of this aggregation
FactTable	Name of fact table to used in this aggregation
FactNumber	Ordinal position of the fact

[0053]

meta—AggregationMeasure Table
AttributeName	Description
AggregationName	Name of aggregation
FactTable	Name of fact table for this aggregation
MeasureName	Name of measure for this aggregation for this fact
	table

[0054]

meta—FactMeasure Table
AttributeName	Description
FactTable	Name of the fact table to which the measure
	applies
MeasureName	Name of the measure
ColumnName	Name of the column for this measure
AggregationFunction	Code for calculating measure value (e.g., count or
	sum)
DataType	Type of measure (e.g., integer)

[0055]

meta—AggregationProperty Table
	AttributeName	Description
	AggregationName	Name of aggregation
	DimName	Name of dimension table
	PropertyName	Name of property within dimension table

[0056]

meta—ReportMeasure Table
AttributeName	Description
AggregationName	Name of aggregation
MeasureName	Name of this measure for report derived from this
	aggregation
MeasureValue	Code for calculating value for this measure in report

[0057]

meta—Fact Table
AttributeName	Description
FactTable	Name of this fact table
FactType	Type of this fact table (e.g., visit and transaction)

[0058] FIG. 16 is a flow diagram of the processing of the aggregation engine in one embodiment. The engine is passed the name of an aggregation and generates the aggregation and report tables as specified by the aggregation metadata. The processing of each of the blocks of this flow diagram is described in detail in a flow chart that follows. In block 1601, the engine creates the aggregation table with the appropriate columns as defined by the aggregation metadata. In block 1602, the function updates the aggregation table to include the aggregating property and measures as defined by the aggregation metadata. In block 1603, the function creates the report table with the appropriate columns as defined by the aggregation metadata. In block 1604, the function updates the report table as defined by the aggregation metadata and then completes.

[0059] FIG. 17 is a flow diagram of the processing of the create aggregation table function in one embodiment. This function is passed the name of an aggregation. This function identifies the aggregation property and each measure from the aggregation metadata that is to be included in the aggregation table for the aggregation. The function then requests the database system to create the table. In block 1701, the function retrieves the entry from the meta_—AggregationProperty A table for the aggregation. In block 1702, the function adds a column for the aggregating property to the aggregation table corresponding to the selected entry. In blocks 1703–1708, the function loops, selecting each measure for each fact table included in the aggregation and adding a corresponding column to the aggregation table. In block 1703, the function selects the next fact table from the meta_—AggregationFact table. In decision block 1704, if all the fact tables of the meta_—AggregationFact table have already been selected, then the function continues at block 1709, else the function continues at block 1705. In block 1705, the function selects the next measure for the selected fact table from the meta_—AggregationMeasure table. In decision block 1706, if all the measures have already been selected for the selected fact table, then the function loops to block 1703 to select the next fact table from the meta_—AggregationFact table, else the function continues at block 1707. In block 1707, the function retrieves the entry from the meta_—FactMeasure table for the selected fact table and the selected measure. In block 1708, the function adds a column to the aggregation table corresponding to the retrieved entry and then loops to block 1705 to select the next measure. In block 1709, the function requests creation of the aggregation table with the added columns.

[0060] FIG. 18 is a flow diagram of the processing of the update aggregation table function in one embodiment. This function is passed the name of the aggregation. The function adds a row to the aggregation table that corresponds to each entry of the dimension table corresponding to the aggregating property. In one embodiment, the function might only add to the aggregation table rows corresponding to rows of the dimension table that are pointed to by the fact tables used in the aggregation. The function then selects each fact table and generates for each row of the aggregation table the measures derived from that fact table. In block 1801, the function retrieves the entry from the meta_—AggregationProperty table. In blocks 1802–1804, the function loops, adding a row to the aggregation table for each row of the dimension table identified by the retrieved entry of the meta_—AggregationProperty table. In block 1802, the function retrieves the next entry from dimension table. In decision block 1803, if all the entries have already been retrieved, then the function continues at block 1805, else the function continues at block 1804. In block 1804, the function adds a row to the aggregation table and then loops to block 1802 to retrieve the next entry from the dimension table. In blocks 1805–1807, the function loops, selecting each fact table that is used in the aggregation and calculating the measures derived from the fact table. In block 1805, the function selects the next fact table from the meta_—AggregationFact table. In decision block 1806, if all the fact tables have already been selected, then the function returns, else the function continues at block 1807. In block 1807, the function aggregates the data of the fact table by invoking the aggregate fact table function and then loops to block 1805 to select the next fact table.

[0061] FIG. 19 is a flow diagram of the processing of the aggregate fact table function in one embodiment. This function is passed the name of an aggregation and the name of a fact table used in that aggregation. The function selects each row of the fact table and aggregates each measure derived from that fact table into the aggregation table. In block 1901, the function selects the next row of the fact table. In decision block 1902, if all the rows of the selected fact table have already been selected, then the function returns, else the function continues at block 1903. In blocks 1903–1906, the function loops, calculating the measure value for the selected row for each measure derived from the fact table. In block 1903, the function selects the next measure from the meta_—AggregationMeasure table. In decision block 1904, if all the measures have already been selected, then the function loops to block 1901 to select the next row of the fact table, else the function continues at block 1905. In block 1905, the function retrieves the entry from the meta_—FactMeasure table. In block 1906, the function updates the selected measure of a row of the aggregation table based on the selected row of the fact table. The updating is in accordance with the aggregation function of the retrieved entry from the meta_—FactMeasure table. The function then loops to block 1903 to select the next measure.

[0062] FIG. 20 is a flow diagram of the processing of the create report table. This function is passed the name of an aggregation. The function creates a report table that includes a column for the aggregating property and a column for each measure in the report. In block 2001, the function retrieves the entry from the meta_—AggregationProperty table for the aggregation. In block 2002, the function adds a column to the report table corresponding to the aggregating property. In blocks 2003–2005, the function loops, adding a column to the report table for each measure to be included in the report. In block 2003, the function selects the next measure from the meta_—ReportMeasure table. In decision block 2004, if all the measures have already been selected, then the function continues at block 2006, else the function continues at block 2005. In block 2005, the function adds a column to the report table corresponding to the selected measure and then loops to block 2003 to select the next measure. In block 2006, the function requests the creation of the report table and then returns.

[0063] FIG. 21 is a flow diagram of the update report table function in one embodiment. This function is passed the name of an aggregation. The function adds a row to the report table corresponding to each row of the aggregation table with measure values calculated in accordance with the meta_—ReportMeasure table. In block 2101, the function selects the next row of the aggregation table. In decision block 2102, if all the rows of the aggregation table have already been selected, then the function returns, else the function continues at block 2103. In block 2103, the function selects the next measure of the meta_—ReportMeasure table. In block 2104, if all the measures have already been selected, then the function continues at block 2106, else the function continues at block 2105. In block 2105, the function calculates the value for this measure using the MeasureValue attribute of the measure and then loops to block 2103 to select the next measure. In block 2106, the function adds a row to the report table corresponding to the measure values calculated in block 2105 and then loops to block 2101 to select the next row is of the aggregation table.

[0064] From the above description, it will be appreciated that although specific embodiments of the technology have been described for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the invention. For example, the import metadata may be used to transform the data of a data store. In such a situation, the data store may be considered to have data that is to be “imported” into other tables of the data store, which may be defined by the metadata itself. Alternatively, the tables defined by the metadata can be part of a separate data store. One skilled in the art will also appreciate that the metadata can be stored in various different ways and need not be stored in separate tables as described above. Accordingly, the invention is not limited except by the appended claims.