TECHNIQUES FOR MODELING AGGREGATION RECORDS

• US 20160275165A1
• Filed: 03/18/2015
• Published: 09/22/2016
• Est. Priority Date: 03/18/2015
• Status: Active Grant

• Current Assignee: Yahoo Assets LLC
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Kilroy, Jonathan David, Nussel, Dale Arthur, Watfa, Allie Khalil, Pardeshi, Mangesh Madan
• Granted Patent: US 10,346,433 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24552 Database cache management
• CPC Class Codes: G06F 16/254 Extract, transform and load [ETL] procedures, e.g. ETL data flows in data warehouses
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP

Abstract

Techniques and technologies described herein can generate and provide identifiers, such as unique identifiers, for individual records of aggregate data. Such identifiers allow systems to cache and reuse individual records of aggregate data. Also, such identifiers can facilitate simpler analytic views of aggregate data.

Claims

1. A method, comprising:

receiving, at a processor, a collection of facts and a multidimensional data structure;

identifying, by the processor, a plurality of records within the collection of facts;

identifying, by the processor, a plurality of dimensions of the multidimensional data structure;

identifying, by the processor, respective values for each dimension of the plurality of dimensions per record of the plurality of records;

concatenating, by the processor, respective values for each dimension of the plurality of dimensions per record of the plurality of records resulting in a composite key per record of the plurality of records; and

directing, by the processor, caching of the composite keys in a cache.

2. The method of claim 1, wherein:

the collection of facts is organized in a fact table associated with online analytics processing (OLAP),

the multidimensional data structure is an OLAP cube,

each dimension of the multidimensional data structure is derived from a dimension table associated with the OLAP cube, and

the processor is an OLAP processor.

3. The method of claim 1, wherein:

the composite key includes a character string including a combination of indications of dimension-value pairs for a record of the plurality of records, and

the concatenation includes concatenating a non-alphanumeric character between each of the indications of dimension-value pairs, so that a key of the composite keys includes the non-alphanumeric character as a separator between each dimension-value pair.

4. The method of claim 1, wherein the concatenation of the respective values for each dimension per record includes dimensions representative of key fields and excludes dimensions representative of metric fields, wherein a metric field of the plurality of dimensions holds a metric (e.g., a measured value) or an averaged metric, and wherein a key field of the plurality of dimensions holds a value that at least partially defines a scope of a metric or an averaged metric.

5. The method of claim 4, wherein a key field of the plurality of dimensions includes a date field, a time field, a product description field, a product region description field, or a product type field, and wherein a metric field of the plurality of dimensions includes a number of impressions, a number of page views, dwell times, or an average of any of the aforementioned metric fields.

6. The method of claim 1, wherein a utility function, executed by the processor, instructs the concatenation of the respective values for each dimension, and wherein the utility function uses an object with dimension and value pairs for a given record to generate the composite key for the given record.

7. The method of claim 1, further comprising rolling up a key of the composite keys by replacing at least one dimension-value pair of the key with a placeholder string indicative of all values for the replaced dimension-value pair, resulting in a rolled-up composite key.

8. The method of claim 7, wherein the rolling up of a key of the composite keys includes:

first rollup concatenating, by the processor, respective values for a first set of dimensions of the plurality of dimensions per record of the plurality of records; and

second rollup concatenating, by the processor, a specific character string to a result of the first rollup concatenation per record, the specific character acting as a placeholder for all values for a second set of dimensions of the plurality of dimensions for all facts of the plurality of records, the second rollup concatenation resulting in a rolled-up composite key per rolled-up record of the plurality of records, each rolled-up composite key including a value for each dimension of the first set of dimensions and the specific character string acting as a placeholder for all values for the second set of dimensions.

9. The method of claim 7, further comprising performing the rolling up of a key for every one of the composite keys in every possible way, resulting in every possible combination of rolled-up composite keys.

10. The method of claim 9, further comprising caching, by the cache, the composite keys and the rolled-up composite keys in the cache, wherein the cache is a block of memory for temporary storage communicatively coupled to the processor, and wherein temporariness of the storage of the cache is relative to the remainder of the memory.

11. Circuitry, comprising:

fact identification circuitry configured to identify respective values for each dimension of a plurality of dimensions of a multidimensional data structure per record of a plurality of records within a collection of facts;

key generation circuitry configured to concatenate the respective values for each dimension of the plurality of dimensions per record of the plurality of records, resulting in a composite key per record of the plurality of records;

key roll-up circuitry configured to roll up the composite keys, resulting in a rolled-up composite key per unique combination of the plurality of dimensions;

cache management circuitry configured to cache the composite keys and the rolled-up keys in a block of memory for temporary storage, wherein temporariness of the storage of the block of memory is relative to the remainder of the memory; and

record retrieval circuitry configured to retrieve a selected record of the collection of facts using a corresponding composite key of the cached composite keys and rolled-up composite keys as a unique identifier for the selected record, wherein the selected record corresponds to one of the cached composite keys or cached rolled-up composite keys.

12. The circuitry of claim 11, wherein the record retrieval circuitry is further configured to bypassing a request for web services corresponding to the retrieval of the selected record of the collection of facts by reusing a corresponding composite key of the cached composite keys and cached rolled-up composite keys as a unique identifier identifying the selected record.

13. The circuitry of claim 11, wherein the record retrieval circuitry is further configured to determining whether the selected record is cached in the cache by identifying whether a corresponding composite key of the cached composite keys and cached rolled-up composite keys is cached in the cache,

wherein if the corresponding composite key is not cached in the cache then the record retrieval circuitry attempts to retrieve the selected record from the collection of facts using a web service, and

wherein if the corresponding composite key is cached in the cache then the record retrieval circuitry attempts to retrieve the selected record from the collection of facts using the corresponding composite key as a unique identifier of the selected record.

14. The circuitry of claim 11, wherein the retrieval circuitry is further configured to receive a selection of a record of the collection of facts via a graphical user interface, prior to the retrieving of the selected record of the collection of facts.

15. The circuitry of claim 11, wherein the retrieval circuitry is further configured to, in retrieving the selected record of the collection of facts, deserialize the composite key using a deserializer.

16. The circuitry of claim 15, wherein the deserializer only recognizes a certain set of dimension-value pairs associated with composite keys.

17. The circuitry of claim 11, further comprising input/output circuitry configured to direct displaying of the retrieved record within a graphical user interface on a display device.

18. The circuitry of claim 17, further comprising display circuitry configured to display the retrieved record within the graphical user interface on the display device.

19. A non-transitory computer readable medium, comprising:

instructions executable by a processor to identify respective values for each dimension of a plurality of dimensions of a multidimensional data structure per record of a plurality of records within a collection of facts;

instructions executable by a processor to concatenate the respective values for each dimension of the plurality of dimensions per record of the plurality of records, resulting in a composite key per record of the plurality of records;

instructions executable by a processor to roll up the composite keys, resulting in a rolled-up composite key per unique combination of the plurality of dimensions;

instructions executable by a processor to cache the composite keys and the rolled-up keys in a block of memory for temporary storage, wherein temporariness of the storage of the block of memory is relative to the remainder of the memory; and

instructions executable by a processor to retrieve a selected record of the collection of facts using a corresponding composite key of the cached composite keys and rolled-up composite keys as a unique identifier for the selected record, wherein the selected record corresponds to one of the cached composite keys or cached rolled-up composite keys.

20. The medium of claim 19, further comprising:

instructions executable by a processor to receive and identify an additional dimension of the multidimensional data structure;

instructions executable by a processor to re-identify respective values for each dimension of the plurality of dimensions, including the additional dimension, per record of the plurality of records;

instructions executable by a processor to re-concatenate respective values for each dimension of the plurality of dimensions, including the additional dimension, per record of the plurality of records resulting in a composite key per record of the plurality of records that account for the additional dimension;

instructions executable by a processor to re-perform the rolling up of every one of the composite keys in every possible way, resulting in every possible combination of rolled-up composite keys, wherein the combinations include the additional dimension; and

instructions executable by a processor to re-cache the composite keys and the rolled-up composite keys in the cache, wherein the composite keys and the rolled-up keys include the additional dimension.