Methods and systems for bulk row save logic in an object relational mapping layer and application framework

• US 8,682,863 B2
• Filed: 02/23/2007
• Issued: 03/25/2014
• Est. Priority Date: 10/04/2006
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Weissman, Craig, Tamm, Steven, Fell, Simon, Wong, Simon, Fisher, Steve, Hansma, Scott
• Primary Examiner(s): KINSAUL, DANIEL W
• Application Number:
  US11/678,477
• Publication Number:
  US 20080086447A1
• Time in Patent Office: 2,587 Days
• Field of Search: None
• US Class Current: 707/684
• CPC Class Codes: G06F 16/21 Design, administration or maintenance of databases
• CPC Class Codes: G06F 16/284 Relational databases

Abstract

In accordance with embodiments, there are provided mechanisms and methods for saving multiple rows together through an object relational mapping layer to a database. These mechanisms and methods for saving multiple rows together can enable embodiments to detect faults in the save operation(s) and recover. The ability of embodiments to detect faults in the save operation(s) and recover can enable embodiments to provide a robust forgiving published API that saves a set of rows together whenever possible.

Claims

1. A method comprising:

receiving, from an object relational mapping layer Application Programming Interface (API) of a database in a known start state, a set of rows to save together in the database, wherein the database comprises a multi-tenant database system having elements of hardware and software shared by multiple customers, and further wherein the set of rows to save together comprises requests for multiple of the customers to be simultaneously processed together in the database;

saving the set of rows together in the database until each row not associated with faulting in the set of rows has been saved in the database;

determining that one or more faults occurred in saving the set of rows;

rolling back the set of rows to the known start state;

rolling back any side effects that were permitted to execute;

rolling back any resource usage counted against governor limits by executing side effects;

forming a new set of rows not associated with the one or more faults to save together in the database from the set of rows by removing rows associated with the one or more faults during the saving of the set of rows in the database;

repeatedly saving in the database and forming using the new set of rows until a set of rows that can be saved in the database from the known start state without fault is determined; and

updating resources used and checking against governor limits.

2. The method of claim 1, wherein receiving, at the object relational mapping layer API in a known start state, a set of rows to save together includes:

receiving information to save to a set of rows in a Simple Object Access Protocol (SOAP) transaction.

3. The method of claim 1, wherein saving the set of rows together until each row in the set of rows has been saved includes:

recording fault information for any row associated with a save operation that faults.

4. The method of claim 3, wherein forming a new set of rows not associated with faulting to save together from the set of rows by removing rows associated with faulting during saving includes:

removing rows corresponding to faults based on the recorded fault information from the set of rows to form the new set of rows to save together.

5. The method of claim 1, further comprising:

committing the set of rows when no fault(s) have occurred.

6. The method of claim 5, wherein committing the set of rows when no fault(s) have occurred comprises ensuring that no side effects have occurred from executing code on behalf of any rows failing to be saved.

7. The method of claim 1, wherein repeatedly saving in the database and forming using the new set of rows until a set of rows that can be saved in the database from the known start state without fault is determined, includes:

stopping triggers from executing;

repeatedly saving in the database and forming using the new set of rows until a set of rows that can be saved in the database from the known start state without fault is determined;

rolling back the set of rows to the known start state;

allowing triggers to execute; and

committing the set of rows that can be saved in the database from the known start state without fault in the database with side effects.

8. The method of claim 7, wherein rolling back the set of rows to the known start state further includes rolling back database governor limits to the known start state.

9. The method of claim 7, wherein the triggers include code for tracking database governor limits.

10. A non-transitory machine-readable medium having instructions stored thereon that, when executed by one or more processors, the instructions cause the one or more processors to perform operations comprising:

receiving, from an object relational mapping layer Application Programming Interface (API) of a database in a known start state, a set of rows to save together in the database, wherein the database comprises a multi-tenant database system having elements of hardware and software shared by multiple customers, and further wherein the set of rows to save together comprises requests for multiple of the customers to be simultaneously processed together in the database;

saving the set of rows together in the database until each row not associated with faulting in the set of rows has been saved in the database;

determining that one or more faults occurred in saving the set of rows;

rolling back the set of rows to the known start state;

rolling back any side effects that were permitted to execute;

rolling back any resource usage counted against governor limits by executing side effects;

forming a new set of rows not associated with the one or more faults to save together in the database from the set of rows by removing rows associated with the one or more faults during the saving of the set of rows in the database;

repeatedly saving in the database and forming using the new set of rows until a set of rows that can be saved in the database from the known start state without fault is determined; and

updating resources used and checking against governor limits.

11. The non-transitory machine-readable medium of claim 10, wherein receiving, at the object relational mapping layer API in a known start state, a set of rows to save together comprises:

receiving information to save to a set of rows in a Simple Object Access Protocol (SOAP) transaction.

12. The non-transitory machine-readable medium of claim 10, wherein saving the set of rows together until each row in the set of rows has been saved comprises:

recording fault information for any row associated with a save operation that faults.

13. The non-transitory machine-readable medium of claim 12, wherein forming a new set of rows not associated with faulting to save together from the set of rows by removing rows associated with faulting during saving comprises:

removing rows corresponding to faults based on the recorded fault information from the set of rows to form the new set of rows not associated with faulting to save together.

14. The non-transitory machine-readable medium of claim 10, wherein the instructions cause the instructions to perform further operations comprising:

committing the set of rows when no fault(s) have occurred.

15. The non-transitory machine-readable medium of claim 14, wherein committing the set of rows when no fault(s) have occurred comprises:

ensuring that no side effects have occurred from executing code on behalf of any rows failing to be saved.

16. An apparatus comprising:

a processor; and

one or more stored sequences of instructions which, when executed by the processor, cause the processor to perform operations comprising;

receiving, from an object relational mapping layer Application Programming Interface (API) of a database in a known start state, a set of rows to save together in the database, wherein the database comprises a multi-tenant database system having elements of hardware and software shared by multiple customers, and further wherein the set of rows to save together comprises requests for multiple of the customers to be simultaneously processed together in the database;

saving the set of rows together in the database until each row not associated with faulting in the set of rows has been saved in the database;

determining that one or more faults occurred in saving the set of rows;

rolling back the set of rows to the known start state;

rolling back any side effects that were permitted to execute;

rolling back any resource usage counted against governor limits by executing side effects;

forming a new set of rows not associated with the one or more faults to save together in the database from the set of rows by removing rows associated with the one or more faults during the saving of the set of rows in the database;

repeatedly saving in the database and forming using the new set of rows until a set of rows that can be saved in the database from the known start state without fault is determined; and

updating resources used and checking against governor limits.

17. A method comprising:

(A) receiving, from an object relational mapping layer Application Programming Interface (API) a database at a rollback point, a set of rows to save together in the database, wherein the database comprises a multi-tenant database system having elements of hardware and software shared by multiple customers, and further wherein the set of rows to save together comprises requests for multiple of the customers to be simultaneously processed together in the database;

(B) attempting to save the set of rows together in the database, while recording fault information for any save operation in the database that faults, until each row not associated with faulting in the set of rows has been attempted to be saved in the database at least once;

(C) determining whether faults occurred in attempting to save the set of rows together in the database;

(D) committing the save in the database of the entire set of rows when the faults are determined not to have occurred in attempting to save the set of rows together in the database;

(i) rolling back the attempting to save the set of rows together in the database to the rollback point when the faults are determined to have occurred in attempting to save the set of rows together in the database;

(ii) rolling back resource limits placed upon code executing due to the attempting to save the set of rows together in the database when the faults are determined to have occurred in attempting to save the set of rows together in the database;

(iii) removing rows corresponding to faults from the set of rows to save together to form a subset of rows not corresponding to faults to save together in the database;

(iv) repeating (B) thorough (D) using the subset of rows as a new set of rows to save together in the database until a subset of rows that can be saved and committed in the database without fault is found, determined; and

updating resources used with respect to resource limits.