System, method and device for index compilation to optimize index usage

Patent metadata

Field	Value
Publication number	US-10635666-B2
Application number	US-201715658028-A
Country	US
Kind code	B2
Filing date	Jul 24, 2017
Priority date	Mar 21, 2015
Publication date	Apr 28, 2020
Grant date	Apr 28, 2020

Abstract

Embodiments relate to a system, method, and device for index compilation to optimize index. Once a CREATE INDEX commands is fired, a first check determines if the user requested has to compile this index. If the index is to be compiled, a corresponding C-file is generated, and a DLL is generated out of the C-file. The DLL is then loaded with server. An INDEX DLL is natively compiled and generated corresponding to each INDEX during definition. The INDEX DLL will contain the required structure and method to perform on INDEX, based on the indexed column data-type.

First claim

What is claimed is:

1. A device comprising:
a transceiver configured to receive a query for querying a database;
a processor configured to specify the query, wherein the query comprises a “CREATE INDEX” statement with a command to create a file by compiling an INDEX, for making data available from the database; and
a memory coupled with the processor, wherein in response to the command to compile, the memory is configured to provide the processor with instructions to natively compile the query, in response to the command, to generate a C-file corresponding to each INDEX in the query, and thereby generate a shared object file, based on the C-file generated, wherein the C-file generated comprises a structure and a mechanism to perform an operation on an INDEX based on an indexed column in the database, wherein the shared object file is a dynamic link library file;
wherein the generated C-file corresponding to each INDEX in the query is specific to a data-type of the indexed column in the database.

2. The device according to claim 1 , wherein the instructions to natively compile the query includes instructions to:
create, based on the query received, the structure corresponding to the index;
generate, based on the query received, an access function for an index tuple value, and a request to find an appropriate slot for an index tuple;
store the structure created, the access function generated, and the access function generated in a buffer, and create the C-file based on the structure, the access function and the request;
compile the C-file created using a pre-stored compiler; and
wherein the memory includes further instructions to instruct the processor to load the C-file along with at least one executable files in the device for an execution of the query.

3. The device according to claim 1 , wherein the shared object file comprises a list of columns in the INDEX, an access function for an index tuple value, or an instruction to find an appropriate slot for an index tuple.

4. The device according to claim 1 , wherein the shared object file comprises a function pointer generated for the INDEX, wherein the function pointer is stored in a relation descriptor table.

5. The device according to claim 1 , wherein the C-file is generated based on the INDEX and the query received, and wherein the shared object file is generated by compiling the C-file generated.

6. The device according to claim 1 , wherein the shared object file is used by an application, or other dynamic link library.

7. The device according to claim 1 , wherein when the query is received, the instructions further comprise instructions to:
check if the INDEX received along with the query is schema bounded, wherein a flag is provided to indicate if tuple belongs to a schema bounded index; and
if the INDEX is schema bounded, utilize a function pointer stored into a corresponding relation descriptor associated with the INDEX, to call the shared object file generated for an execution of the query.

8. The device as claimed in in claim 1 , wherein the query is received in a format comprising CREATE INDEX query along with [COMPILE] option.

9. A method comprising:
receiving, by a transceiver, a query for querying a database for making data available from the database;
identifying, by a processor, the query, wherein the query comprises a “CREATE INDEX” statement with a command to create a file by compiling the INDEX; and
natively compiling the query, by the processor, in response to the command, to generate a C-file corresponding to each INDEX in the query, and thereby generate a shared object file, based on the C-file generated, wherein the C-file comprises a structure and a mechanism to perform an operation on an INDEX based on an indexed column in the database, and wherein the shared object file is a dynamic link library file;
wherein the generated C-file corresponding to each INDEX in the query is specific to a data-type of the indexed column in the database.

10. The method according to claim 9 , wherein natively compiling the query comprises:
creating, by the processor, based on the query received, the structure corresponding to the index;
generating, by the processor, based on the query received, specifically based on an INDEX definition given as part of CREATE INDEX command, an access function for an index tuple value, and an instruction to find an appropriate slot for an index tuple, wherein a number of access functions generated for each INDEX are always same;
storing, by the processor, the structure created, the access function generated, and the access function generated in a buffer;
creating, by the processor, the C-file based on the structure, the access function and the instruction;
compiling, by the processor, the C-file created using a pre-stored compiler; and
loading, by the processor, the C-file along with an executable file for an execution of the query.

11. The method according to claim 9 , wherein the shared object file comprises a list of columns in the INDEX, an access function for an index tuple value, or an instruction to find an appropriate slot for an index tuple.

12. The method according to claim 9 , wherein natively compiling the query comprises: creating, by the processor, a system table for the INDEX, and thereby storing an option for compiling the INDEX, wherein the option is selected from a “true” if the command to create the file is present in the query or “false” if the command to create the file is absent from the query.

13. The method according to claim 9 , wherein the shared object file comprises a function pointer generated for the INDEX, wherein the function pointer generated is stored in a relation descriptor table.

14. The method according to claim 9 , wherein natively compiling the query comprises:
when the query has a command to compile the INDEX, generating the shared object file by compiling the C-file generated.

15. The method according to claim 9 , wherein the shared object file is used by an application or other dynamic link library.

16. The method according to claim 9 , wherein when the query is received, the method comprises:
checking, by the processor, if the INDEX received along with the query is schema bounded, wherein a flag is provided to indicate if a tuple belongs to a schema bounded index; and
if the INDEX is schema bounded, utilizing, by the processor, a function pointer stored into a corresponding relation descriptor associated with the INDEX, to call the shared object file generated for an execution of the query.

17. The method according to claim 9 , wherein the query is received in a format comprising CREATE INDEX query along with [COMPILE] option.

Assignees

• Huawei Tech Co Ltd

Inventors

• Rastogi Kumar Rajeev
• Kumar Dilip
• Sreekantaiah Nirmala

Classifications

• G06F16/2453 Primary

  Query optimisation · CPC title

• G06F16/2282

  Tablespace storage structures; Management thereof · CPC title

• G06F8/41

  Compilation · CPC title

• G06F9/44521

  Dynamic linking or loading; Link editing at or after load time, e.g. Java class loading · CPC title

• G06F16/2228 Primary

  Indexing structures · CPC title