Relational database schema version management

• US 7,562,357 B2
• Filed: 03/08/2004
• Issued: 07/14/2009
• Est. Priority Date: 03/08/2004
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Vitaldevara, Krishna, Tamma, Leela S.
• Primary Examiner(s): Zhen; Wei Y
• Assistant Examiner(s): Chen; Qing
• Application Number:
  US10/796,613
• Publication Number:
  US 20050198630A1
• Time in Patent Office: 1,954 Days
• Field of Search: 717120-123, 717168-178, 707/200, 707/203
• US Class Current: 717/170
• CPC Class Codes: G06F 16/213 with details for schema evolution support
• CPC Class Codes: G06F 16/29 Geographical information databases
• CPC Class Codes: G06F 8/65 Updates security arrangements therefor G06F21/57
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

Techniques for managing multiple versions of relational database schema data are described. Schema data associated with multiple versions of a relational database are maintained according to a database schema version management structure. Laws of set theory are applied to the structured database schema data to identify data definition language (DDL) and data manipulation language (DML) scripts that are associated with a particular version of the relational database. The identified scripts are then copied into an installation file that can be executed to create a full install of the particular version of the relational database. Similarly, laws of set theory are applied to the structured database schema data to identify DDL scripts, DML scripts, and drop scripts that are to be applied to upgrade from one version of the relational database to a newer version of the relational database.

Claims

1. A method for generating an installation file for a particular version of a relational database, the method comprising:

given a database system of a particular version, the particular version being unknown prior to execution of the method, determining the particular version of the relational database wherein metadata exists that describes a sequence of multiple versions of the relational database where each version is an upgrade from a previous version, and the particular version is not a first version in the sequence;

automatically determining a first set of scripts, the first set of scripts comprising data definition language (DDL) scripts associated with implementing the particular version of the relational database, wherein the first set of scripts are selected from a plurality of scripts, one or more of the plurality of scripts being not associated with implementing the particular version;

the automatically determining a first set of scripts comprising;

extracting a set A₁ comprising one or more filenames from metadata associated with a first version in the sequence, the one or more filenames associated with a file comprising a data definition language (DDL) script associated with the first version;

iteratively extracting a set A_i comprising zero or more filenames from metadata associated with an i^th version of the relational database, the zero or more filenames each associated with a file comprising a data definition language script to be executed when upgrading from version i−

1 of the relational database to version i of the relational database, where i varies incrementally from 2 to j, where the particular version is j; and

determining the first set of scripts as a set theory union of sets A₁, A₂, . . . , A_j;

automatically determining a second set of scripts, the second set of scripts comprising data manipulation language (DML) scripts associated with implementing the particular version of the relational database, wherein the second set of scripts are selected from a plurality of scripts, one or more of the plurality of scripts being not associated with implementing the particular version;

the automatically determining a second set of scripts comprising;

extracting a set A₁ comprising one or more filenames from metadata associated with a first version in the sequence, the one or more filenames associated with a file comprising a data manipulation language (DML) script associated with the first version;

iteratively extracting a set A_i comprising zero or more filenames from metadata associated with an i^th version of the relational database, the zero or more filenames each associated with a file comprising a data manipulation language script to be executed to add or modify a DML object when upgrading from version i−

1 of the relational database to version i of the relational database, where i varies incrementally from 2 to j, where the particular version is j;

iteratively extracting a set B_i comprising zero or more filenames from metadata associated with an i^th version of the relational database, the zero or more filenames each associated with a file comprising a DML drop script to be executed to drop a DML object when upgrading from version i−

1 of the relational database to version i of the relational database, where i varies incrementally from 2 to j, where the particular version is j; and

determining the second set of scripts C_j by determining;

C₂=[A₁∪

A₂]−

B₂;

C₃=[C₂∪

A₃]−

B₃;

C₄=[C₃∪

A₄]−

B₄;

. . .

C_j=[C<sub>j−

1</sub>∪

A_j]−

B_j;

and

generating an installation file comprising a set theory union of the first set of scripts and the second set of scripts.

2. The method as recited in claim 1 wherein the particular version is associated with a first version in a sequence of one or more versions of the relational database.

3. The method as recited in claim 1 wherein the automatically determining a first set comprises extracting a filename from metadata associated with the particular version, the filename associated with a file comprising a data definition language script.

4. The method as recited in claim 1 wherein the automatically determining a second set comprises extracting a filename from metadata associated with the particular version, the filename associated with a file comprising a data manipulation language script.

5. The method as recited in claim 1 wherein the generating an installation file comprises copying a data definition language script from a script file associated with the first set into the installation file.

6. The method as recited in claim 1 wherein the generating an installation file comprises copying a data manipulation language script from a script file associated with the second set into the installation file.

7. The method as recited in claim 6 wherein the copying further comprises prepending a create command to the data manipulation language script in the installation file.

8. The method as recited in claim 1 wherein the metadata comprises an XML file.

9. One or more computer-readable media having computer-readable instructions recorded thereon which, when executed by a computer, cause the computer to implement the method as recited in claim 1.

10. A method for generating an upgrade file to upgrade version i of a relational database to version j of the relational database, where j>

i, the method comprising;

determining a set A of data definition language (DDL) scripts that, when executed, perform creates, alters, and drops of DDL objects associated with version i of the relational database, resulting in DDL objects associated with version j of the relational database, wherein the determining a set A comprises;

iteratively extracting sets M_k, each comprising zero or more filenames from metadata associated with a k^th version of the relational database, where i<

k<

=j, the zero or more filenames each associated with a file comprising a data definition language script to be executed when upgrading from version k−

1 of the relational database to version k of the relational database; and

determining the set A as the set theory union of sets M_i+1, M_i+2, . . . , M_j (A=M_i+1∪

M_i+2∪

. . . ∪

M_j);

determining a set B of data manipulation language (DML) scripts that, when executed, create DML objects that are associated with version j of the relational database, but that are not associated with version i of the relational database, wherein the determining a set B comprises;

determining a set E of DML scripts that when executed;

perform alters DML objects associated with version i and version j of the relational database, but that differ between version i and version j of the relational database; and

perform creates of DML objects that are associated with version j of the relational database but that are not associated with version i of the relational database; and

determining set B as the difference between sets E and C_(B=E−

C);

determining a set C of DML scripts that, when executed, modify DML objects that are associated with both version i and version j of the relational database, but that differ between version i and version j of the relational database, wherein the determining a set C comprises;

determining a set E of DML scripts that when executed;

perform alters of DML objects associated with version i and version j of the relational database, but that differ between version i and version j of the relational database; and

perform creates of DML objects that are associated with version j of the relational database but that are not associated with version i of the relational database;

determining a set F_j of DML scripts that when executed, create DML objects associated with version j of the relational database; and

determining set C as a set theory intersection of set E and set F_j (C=E∩

F_j);

determining a set D of DML drop scripts that, when executed, drop DML objects that are associated with version i of the relational database, but that are not associated with version j of the relational database, wherein the determining a set D comprises;

determining a set E of DML scripts that when executed;

perform alters of DML objects associated with version i and version j of the relational database, but that differ between version i and version j of the relational database; and

perform creates of DML objects that are associated with version j of the relational database but that are not associated with version i of the relational database;

iteratively determining a set F_x of DML scripts that when executed, drop DML objects associated with version x−

1 of the relational database that are not associated with version x of the relational database, where x varies incrementally from i+1 to j;

determining a set G as the set theory union of sets F_i, F_i+1, F_i+2, . . . , F_j (G=F_i∪

F_i+1∪

. . . ∪

F_i+2); and

determining set D as the difference between set G and set E_(D=G−

E); and

generating an upgrade file comprising a set theory union of sets A, B, C, and D (A∪

B∪

C∪

D).

11. The method as recited in claim 10 wherein the determining a set E comprises:

iteratively determining a set P_x of DML scripts that when executed will upgrade DML objects from version x−

1 of the relational database to version x of the relational database, where x varies incrementally from i+1 to j ;

iteratively determining a set N_x of DML scripts that when executed will drop DML objects that are associated with version x−

1 of the relational database but that are not associated with version x of the relational database, where x varies incrementally from i+2 to j ;

iteratively determining a set M_x of DML scripts that when executed will upgrade DML objects from version i of the relational database to version x of the relational database, where x varies incrementally from i+1 to j, and where;

M_i+1=P_i+1

M_i+2=[M_i+1∪

P_i+2]−

N_i+2

M_i+3=[M_i+2∪

P_i+3]−

N_i+3

. . .

M_j=[M<sub>j−

1</sub>∪

P_j]−

N_j; and

determining set E=M_j.

12. The method as recited in claim 10 wherein the determining a set F_j comprises:

extracting a set M₁ comprising one or more filenames from metadata associated with a first version in a sequence of multiple versions of the relational database, the one or more filenames associated with a file comprising a data manipulation language (DML) script associated with the first version;

iteratively extracting a set M_x comprising zero or more filenames from metadata associated with version x of the relational database, the zero or more filenames each associated with a file comprising a DML script to be executed to add or modify a DML object when upgrading from version x−

1 of the relational database to version x of the relational database, where x varies incrementally from 2 to j;

iteratively extracting a set B_x comprising zero or more filenames from metadata associated with version x of the relational database, the zero or more filenames each associated with a file comprising a DML drop script to be executed to drop a DML object when upgrading from version x−

1 of the relational database to version x of the relational database, where x varies incrementally from 2 to j; and

determining the set F_j by determining;

F₂=[M₁∪

M₂]−

B₂,

F₃=[F₂∪

M₃]−

B₃,

F₄=[F₃∪

M₄]−

B₄,

. . .

F_j=[F<sub>j−

1</sub>∪

M_j]−

B_j.