MacLochlainns Weblog

Last week, I wrote about how to use bash arrays and the MySQL database to create unit and integration test scripts. While the MySQL example was nice for some users, there were some others who wanted me to show how to write bash shell scripts for Oracle unit and integration testing. That’s what this blog post does.

If you don’t know much about bash shell, you should start with the prior post to learn about bash arrays, if-statements, and for-loops. In this blog post I only cover how to implement a bash shell script that runs SQL scripts in silent mode and then queries the database in silent mode and writes the output to an external file.

I’ve copied the basic ERD for the example because of a request from a reader. In their opinion it makes cross referencing the two posts unnecessary.

To run the bash shell script, you’ll need the following SQL files, which you can see by clicking not he title below. There are several differences. For example, Oracle doesn’t support a DROP IF EXISTS syntax and requires you to write anonymous blocks in their PL/SQL language; and you must explicitly issue a QUIT; statement even when running in silent mode unlike MySQL, which implicitly issues an exit.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30

-- Drop actor table and actor_s sequence.
BEGIN
  FOR i IN (SELECT   object_name
            ,        object_type
            FROM     user_objects
            WHERE    object_name IN ('ACTOR','ACTOR_S')) LOOP
    IF    i.object_type = 'TABLE' THEN
      EXECUTE IMMEDIATE 'DROP TABLE ' || i.object_name || ' CASCADE CONSTRAINTS';
    ELSIF i.object_type = 'SEQUENCE' THEN
      EXECUTE IMMEDIATE 'DROP SEQUENCE ' || i.object_name;
    END IF;
  END LOOP;
END;
/
 
-- Create an actor table.
CREATE TABLE actor
( actor_id    NUMBER CONSTRAINT actor_pk PRIMARY KEY
, actor_name  VARCHAR(30)  NOT NULL );
 
-- Create an actor_s sequence.
CREATE SEQUENCE actor_s;
 
-- Insert two rows.
INSERT INTO actor VALUES (actor_s.NEXTVAL,'Chris Hemsworth');
INSERT INTO actor VALUES (actor_s.NEXTVAL,'Chris Pine');
INSERT INTO actor VALUES (actor_s.NEXTVAL,'Chris Pratt');
 
-- Quit session.
QUIT;

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32

-- Drop film table and film_s sequence.
BEGIN
  FOR i IN (SELECT   object_name
            ,        object_type
            FROM     user_objects
            WHERE    object_name IN ('FILM','FILM_S')) LOOP
    IF    i.object_type = 'TABLE' THEN
      EXECUTE IMMEDIATE 'DROP TABLE ' || i.object_name || ' CASCADE CONSTRAINTS';
    ELSIF i.object_type = 'SEQUENCE' THEN
      EXECUTE IMMEDIATE 'DROP SEQUENCE ' || i.object_name;
    END IF;
  END LOOP;
END;
/
 
-- Create a film table.
CREATE TABLE film
( film_id    NUMBER CONSTRAINT film_pk PRIMARY KEY
, film_name  VARCHAR(30)  NOT NULL );
 
-- Create an actor_s sequence.
CREATE SEQUENCE film_s;
 
-- Insert four rows.
INSERT INTO film VALUES (film_s.NEXTVAL,'Thor');
INSERT INTO film VALUES (film_s.NEXTVAL,'Thor: The Dark World');
INSERT INTO film VALUES (film_s.NEXTVAL,'Star Trek');
INSERT INTO film VALUES (film_s.NEXTVAL,'Star Trek into Darkness');
INSERT INTO film VALUES (film_s.NEXTVAL,'Guardians of the Galaxy');
 
-- Quit session.
QUIT;

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81

-- Drop movie table and movie_s sequence.
BEGIN
  FOR i IN (SELECT   object_name
            ,        object_type
            FROM     user_objects
            WHERE    object_name IN ('MOVIE','MOVIE_S')) LOOP
    IF    i.object_type = 'TABLE' THEN
      EXECUTE IMMEDIATE 'DROP TABLE ' || i.object_name || ' CASCADE CONSTRAINTS';
    ELSIF i.object_type = 'SEQUENCE' THEN
      EXECUTE IMMEDIATE 'DROP SEQUENCE ' || i.object_name;
    END IF;
  END LOOP;
END;
/
 
-- Create an movie table.
CREATE TABLE movie
( movie_id   NUMBER  CONSTRAINT movie_pk   PRIMARY KEY
, actor_id   NUMBER  CONSTRAINT movie_nn1  NOT NULL
, film_id    NUMBER  CONSTRAINT movie_nn2  NOT NULL
, CONSTRAINT actor_fk FOREIGN KEY (actor_id)
  REFERENCES actor (actor_id)
, CONSTRAINT film_fk  FOREIGN KEY (film_id)
  REFERENCES film(film_id));
 
-- Create table constraint.
CREATE SEQUENCE movie_s;
 
-- Insert translation rows.
INSERT INTO movie
VALUES
( movie_s.NEXTVAL
,(SELECT   actor_id
  FROM     actor
  WHERE    actor_name = 'Chris Hemsworth')
,(SELECT   film_id
  FROM     film
  WHERE    film_name = 'Thor'));
 
INSERT INTO movie
VALUES
( movie_s.NEXTVAL
,(SELECT   actor_id
  FROM     actor
  WHERE    actor_name = 'Chris Hemsworth')
,(SELECT   film_id
  FROM     film
  WHERE    film_name = 'Thor: The Dark World'));
 
INSERT INTO movie
VALUES
( movie_s.NEXTVAL
,(SELECT   actor_id
  FROM     actor
  WHERE    actor_name = 'Chris Pine')
,(SELECT   film_id
  FROM     film
  WHERE    film_name = 'Star Trek'));
 
INSERT INTO movie
VALUES
( movie_s.NEXTVAL
,(SELECT   actor_id
  FROM     actor
  WHERE    actor_name = 'Chris Pine')
,(SELECT   film_id
  FROM     film
  WHERE    film_name = 'Star Trek into Darkness'));
 
INSERT INTO movie
VALUES
( movie_s.NEXTVAL
,(SELECT   actor_id
  FROM     actor
  WHERE    actor_name = 'Chris Pratt')
,(SELECT   film_id
  FROM     film
  WHERE    film_name = 'Guardians of the Galaxy'));
 
-- Quit session.
QUIT;

1
2
3
4
5
6
7
8
9

-- Set Oracle column width.
COL table_name FORMAT A30 HEADING "Table Name"
 
-- Query the tables.
SELECT   table_name
FROM     user_tables;
 
-- Exit SQL*Plus.
QUIT;

1
2
3
4
5
6
7
8
9
10
11

-- Format query.
COL film_actors FORMAT A40 HEADING "Actors in Films"
 
-- Diagnostic query.
SELECT   a.actor_name || ', ' || f.film_name AS film_actors
FROM     actor a INNER JOIN movie m
ON       a.actor_id = m.actor_id INNER JOIN film f
ON       m.film_id = f.film_id;
 
-- Quit the session.
QUIT;

If you don’t have a sample test schema to use to test this script, you can create a sample schema with the following create_user.sql file. The file depends on the existence of a users and temp tablespace.

Click the link below to see the source code for a script that let’s you create a sample user account as the system user:

1
2

SET LONG 200000
SELECT dbms_metadata.get_ddl('TABLESPACE','USERS') FROM dual;

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

-- Drop the sample user table.
BEGIN
  FOR i IN (SELECT   username
            FROM     dba_users
            WHERE    username = 'SAMPLE') LOOP
    EXECUTE IMMEDIATE 'DROP USER ' || i.username || ' CASCADE';
  END LOOP;
END;
/
 
-- Create the sample user table.
CREATE USER sample IDENTIFIED BY sample
  DEFAULT TABLESPACE users
  TEMPORARY TABLESPACE temp
  QUOTA 50M ON users;
 
-- Grant privileges to sample user.
GRANT CREATE CLUSTER, CREATE INDEXTYPE, CREATE OPERATOR
,     CREATE PROCEDURE, CREATE SEQUENCE, CREATE SESSION
,     CREATE TABLE, CREATE TRIGGER, CREATE TYPE
,     CREATE VIEW TO sample;

The following list_oracle.sh shell script expects to receive the username, password, and fully qualified path in that specific order. The script names are entered manually in the array because this should be a unit test script.

This is an insecure version of the list_oracle.sh script because you provide the password on the command line. It’s better to provide the password as you run the script.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

#!/usr/bin/bash
 
# Assign user and password
username="${1}"
password="${2}"
directory="${3}"
 
echo "User name:" ${username}
echo "Password: " ${password}
echo "Directory:" ${directory}
 
# Define an array.
declare -a cmd
 
# Assign elements to an array.
cmd[0]="actor.sql"
cmd[1]="film.sql"
cmd[2]="movie.sql"
 
# Call the array elements.
for i in ${cmd[*]}; do
  sqlplus -s ${username}/${password} @${directory}/${i} > /dev/null
done
 
# Connect and pipe the query result minus errors and warnings to the while loop.
sqlplus -s ${username}/${password} @${directory}/tables.sql 2>/dev/null |
 
# Read through the piped result until it's empty.
while IFS='\n' read actor_name; do
  echo $actor_name
done
 
# Connect and pipe the query result minus errors and warnings to the while loop.
sqlplus -s ${username}/${password} @${directory}/result.sql 2>/dev/null |
 
# Read through the piped result until it's empty.
while IFS='\n' read actor_name; do
  echo $actor_name
done

The IFS (Internal Field Separator) works with whitespace by default. The IFS on lines 29 and 37 sets the IFS to a line return ('\n'). That’s the trick to display the data, and you can read more about the IFS in this question and answer post.

You can run the shell script with the following syntax:

./list_oracle.sh sample sample /home/student/Code/bash/oracle > output.txt

You can then display the results from the output.txt file with the following command:

cat output.txt command:

It will display the following output:

User name: sample
Password:  sample
Directory: /home/student/Code/bash/oracle
 
Table Name
------------------------------
MOVIE
FILM
ACTOR
 
Actors in Films
----------------------------------------
Chris Hemsworth, Thor
Chris Hemsworth, Thor: The Dark World
Chris Pine, Star Trek
Chris Pine, Star Trek into Darkness
Chris Pratt, Guardians of the Galaxy

As always, I hope this helps those looking for a solution.

This is the continuation of my efforts to stage an awesome Fedora developer’s instance. It shows you how to install Java 1.8 software development kit, which is nice to have. Though you can’t use Java 1.8 officially with Oracle SQL Developer 4.0.3 it is required for Oracle SQL Developer 4.1. Fortunately, the Oracle Product Manager, Jeff Smith has advised us that you can use Java 1.8 JDK with Oracle SQL Developer 4.0.3, and he’s written a comment to the blog post that it runs better with the Java 1.8 SDK.

After you install Oracle SQL Developer 4.0.3 or Oracle SQL Developer 4.1, you can watch Jeff Smith’s YouTube Video on SQL Developer 3.1 to learn how to use the basics of SQL Developer. I couldn’t find an updated version of the video for SQL Developer 4 but I didn’t try too hard.

You use yum as the root user to install Java SDK 1.8, much like my earlier Installing the Java SDK 1.7 and Java-MySQL Sample Program. The following command installs Java 8:

yum install -y java-1.8*

It produces the following output:

Loaded plugins: langpacks, refresh-packagekit
fedora/20/x86_64/metalink                                   |  18 kB  00:00     
mysql-connectors-community                                  | 2.5 kB  00:00     
mysql-tools-community                                       | 2.5 kB  00:00     
mysql56-community                                           | 2.5 kB  00:00     
pgdg93                                                      | 3.6 kB  00:00     
updates/20/x86_64/metalink                                  |  16 kB  00:00     
updates                                                     | 4.9 kB  00:00     
(1/2): mysql-tools-community/20/x86_64/primary_db           |  21 kB  00:00     
(2/2): updates/20/x86_64/primary_db                         |  13 MB  00:09     
updates/20/x86_64/pkgtags
updates
(1/2): updates/20/x86_64/pkgtags                            | 1.4 MB  00:02     
(2/2): updates/20/x86_64/updateinfo                         | 1.9 MB  00:04     
Package 1:java-1.8.0-openjdk-headless-1.8.0.31-1.b13.fc20.x86_64 already installed and latest version
Package 1:java-1.8.0-openjdk-javadoc-1.8.0.31-1.b13.fc20.noarch already installed and latest version
Resolving Dependencies
--> Running transaction check
---> Package java-1.8.0-openjdk.x86_64 1:1.8.0.31-1.b13.fc20 will be installed
---> Package java-1.8.0-openjdk-accessibility.x86_64 1:1.8.0.31-1.b13.fc20 will be installed
---> Package java-1.8.0-openjdk-demo.x86_64 1:1.8.0.31-1.b13.fc20 will be installed
---> Package java-1.8.0-openjdk-devel.x86_64 1:1.8.0.31-1.b13.fc20 will be installed
---> Package java-1.8.0-openjdk-src.x86_64 1:1.8.0.31-1.b13.fc20 will be installed
--> Finished Dependency Resolution
 
Dependencies Resolved
 
================================================================================
 Package                          Arch   Version                  Repository
                                                                           Size
================================================================================
Installing:
 java-1.8.0-openjdk               x86_64 1:1.8.0.31-1.b13.fc20    updates 201 k
 java-1.8.0-openjdk-accessibility x86_64 1:1.8.0.31-1.b13.fc20    updates  12 k
 java-1.8.0-openjdk-demo          x86_64 1:1.8.0.31-1.b13.fc20    updates 1.9 M
 java-1.8.0-openjdk-devel         x86_64 1:1.8.0.31-1.b13.fc20    updates 9.2 M
 java-1.8.0-openjdk-src           x86_64 1:1.8.0.31-1.b13.fc20    updates  45 M
 
Transaction Summary
================================================================================
Install  5 Packages
 
Total download size: 56 M
Installed size: 92 M
Downloading packages:
(1/5): java-1.8.0-openjdk-accessibility-1.8.0.31-1.b13.fc20 |  12 kB  00:00     
(2/5): java-1.8.0-openjdk-1.8.0.31-1.b13.fc20.x86_64.rpm    | 201 kB  00:02     
(3/5): java-1.8.0-openjdk-demo-1.8.0.31-1.b13.fc20.x86_64.r | 1.9 MB  00:03     
(4/5): java-1.8.0-openjdk-devel-1.8.0.31-1.b13.fc20.x86_64. | 9.2 MB  00:07     
(5/5): java-1.8.0-openjdk-src-1.8.0.31-1.b13.fc20.x86_64.rp |  45 MB  05:05     
--------------------------------------------------------------------------------
Total                                              187 kB/s |  56 MB  05:05     
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction (shutdown inhibited)
  Installing : 1:java-1.8.0-openjdk-1.8.0.31-1.b13.fc20.x86_64              1/5 
  Installing : 1:java-1.8.0-openjdk-devel-1.8.0.31-1.b13.fc20.x86_64        2/5 
  Installing : 1:java-1.8.0-openjdk-demo-1.8.0.31-1.b13.fc20.x86_64         3/5 
  Installing : 1:java-1.8.0-openjdk-accessibility-1.8.0.31-1.b13.fc20.x86   4/5 
  Installing : 1:java-1.8.0-openjdk-src-1.8.0.31-1.b13.fc20.x86_64          5/5 
  Verifying  : 1:java-1.8.0-openjdk-devel-1.8.0.31-1.b13.fc20.x86_64        1/5 
  Verifying  : 1:java-1.8.0-openjdk-demo-1.8.0.31-1.b13.fc20.x86_64         2/5 
  Verifying  : 1:java-1.8.0-openjdk-1.8.0.31-1.b13.fc20.x86_64              3/5 
  Verifying  : 1:java-1.8.0-openjdk-accessibility-1.8.0.31-1.b13.fc20.x86   4/5 
  Verifying  : 1:java-1.8.0-openjdk-src-1.8.0.31-1.b13.fc20.x86_64          5/5 
 
Installed:
  java-1.8.0-openjdk.x86_64 1:1.8.0.31-1.b13.fc20                               
  java-1.8.0-openjdk-accessibility.x86_64 1:1.8.0.31-1.b13.fc20                 
  java-1.8.0-openjdk-demo.x86_64 1:1.8.0.31-1.b13.fc20                          
  java-1.8.0-openjdk-devel.x86_64 1:1.8.0.31-1.b13.fc20                         
  java-1.8.0-openjdk-src.x86_64 1:1.8.0.31-1.b13.fc20                           
 
Complete!

Then, you go to Oracle’s SQL Developer 4.0.3 web page or Oracle’s Beta SQL Developer 4.1 web page and download the SQL Developer RPM. At the time of writing, you download the following SQL Developer 4.0.3 RPM:

sqldeveloper-4.0.3.16.84-1.noarch.rpm

Assuming you download the sqldeveloper-4.0.3.16.84-1.noarch.rpm file to the student user’s account. It will download into the /home/student/Downloads directory. You run the SQL Developer RPM file with the following syntax as the root user:

rpm -Uhv /home/student/Downloads/sqldeveloper-4.0.3.16.84-1.noarch.rpm

Running the SQL Developer RPM produces the following output:

Preparing...                          ################################# [100%]
Updating / installing...
   1:sqldeveloper-4.0.3.16.84-1       ################################# [100%]

You can now run the sqldeveloper.sh file as the root user with the following syntax:

/opt/sqldeveloper/sqldeveloper.sh

At this point, it’s important to note that my download from the Oracle SQL Developer 4.1 page turned out to be SQL Developer 4.0.3. It prompts you for the correct Java JDK, as shown below. You may opt to enter the path to the Java JDK 1.8 for SQL Developer 4.1 because until today you downloaded the Oracle SQL Developer 4.0.3 version from the Oracle SQL Developer 4.1 page. Naturally, the Oracle SQL Developer 4.1 instructions say to use the Java 1.8 JDK on the RPM for Linux Installation Notes web page, as shown below:

If you assume from the instructions on the Oracle instruction page above that Oracle SQL Developer 4.0.3 and Oracle SQL Developer 4.1 support Java 1.8 JDK, you may enter the location for the Java JDK 1.8 when prompted. Jeff Smith, the Product Manager wrote this blog post on Oracle SQL Developer 4: Windows and the JDK. Unfortunately, you’ll see the following message if you attempt to run Oracle SQL Developer 4.0.3 with the Java 1.8 SDK at the command-line:

 Oracle SQL Developer
 Copyright (c) 1997, 2014, Oracle and/or its affiliates. All rights reserved.
 
Type the full pathname of a JDK installation (or Ctrl-C to quit), the path will be stored in /root/.sqldeveloper/4.0.0/product.conf
/usr/lib/jvm/java-1.8.0-openjdk-1.8.0.31.x86_64
OpenJDK 64-Bit Server VM warning: ignoring option MaxPermSize=256M; support was removed in 8.0

It also raises the following error message dialog:

Text version of Unsupported JDK Version error message:

The error dialog message tells us that the instructions on the RPM for Linux Installation Notes web page can be misleading. You really need to use the Java JDK 1.7 to be supported officially, but you can safely ignore the error.

If you want a certified component, leave the “Skip This Message Next Time” checkbox unchecked and click the “No” button to continue. At this point, there’s no automatic recovery. You need to open the following file:

/root/.sqldeveloper/4.0.0/product.conf

You need to change the SetJavaHome parameter in the file to the following:

# SetJavaHome /path/jdk
SetJavaHome /usr/lib/jvm/java-1.7.0-openjdk-1.7.0.79-2.5.5.0.fc20.x86_64

After making the change, you can re-run the sqldeveloper.sh shell as follows:

/opt/sqldeveloper/sqldeveloper.sh

It launches the following dialog message:

The installation pauses to ask you if you want to transfer an existing SQL Developer configuration by raising the following dialog. Assuming this is a new installation, the installer won’t find a prior configuration file. You need to click the “No” button to proceed.

The installation continues and launches SQL Developer. The first time launch shows you the following Oracle Usage Tracking dialog. If you don’t want your use monitored, uncheck the “Allow automated usage reporting to Oracle” checkbox. Click the “OK” button to continue.

After dismissing the Oracle Usage Tracking dialog, you see the SQL Developer environment:

After installing SQL Developer in the root account, you can install it as the student user. You use this command as the student user:

/opt/sqldeveloper/sqldeveloper.sh

It returns the following error because it’s the second installation and SQL Developer doesn’t prompt you to configure the user’s product.conf file with the working JDK location:

 Oracle SQL Developer
 Copyright (c) 1997, 2014, Oracle and/or its affiliates. All rights reserved.
 
Type the full pathname of a JDK installation (or Ctrl-C to quit), the path will be stored in /home/student/.sqldeveloper/4.0.0/product.conf
Error:  Unable to get APP_JAVA_HOME input from stdin after 10 tries

You need to edit the /home/student/.sqldeveloper/4.0.0/product.conf file, and add the following line to the file:

# SetJavaHome /path/jdk
SetJavaHome /usr/lib/jvm/java-1.7.0-openjdk-1.7.0.79-2.5.5.0.fc20.x86_64

Now, you can launch SQL Developer with the following command:

/opt/sqldeveloper/sqldeveloper.sh

Alternatively, you can add the following alias to the student user’s .bashrc file:

# Set alias for SQL Developer tool.
alias sqldeveloper="/opt/sqldeveloper/sqldeveloper.sh"

You can now launch the SQL Developer tool, like this as the student user:

sqldeveloper

You see the following when SQL Developer launches:

As always, I hope this helps those trying to sort out installing SQL Developer on a Fedora server.

The following walks you through how you sign on to a STUDENT Workspace with Oracle’s APEX product. It shows you how to create a new table with the Object Browser tool.

You can find instructions on how to create your own STUDENT Workspace in this blog post. Overall, Oracle APEX is a valuable tool to learn and master.

1. You start the process by accessing the Oracle Database 11g APEX, which you can access at http://localhost:8080/apex by default on the server. If you’ve got a static IP address for your instance, you can replace localhost with the IP address or hostname for the IP address.
   • Workspace: STUDENT
   • Username:  ADMIN
   • Password:  STUDENT

2. After you login to the STUDENT workspace, you have four options. They are the: Application Builder, SQL Workshop, Team Development, and Administration. You start the process by accessing the Oracle Database 11g APEX, which you can access at http://localhost:8080/apex by default on the server. If you’ve got a static IP address for your instance, you can replace localhost with the IP address or hostname for the IP address. Click on the Object Browser icon to proceed.

3. Clicking the SQL Workshop icon takes you to the second level menu. You click the Object Browser icon to create a database object.

4. After clicking the Object Browser icon, you see the screen at the left. Click the Create button to create a table.

5. After clicking the Create button, you see the screen at the left. Click the type of database object that you want to create. In our case, we click the Table hypertext to start the create table workflow.

6. After clicking the Table hyperlink, you see the Create Table screen at the left. Enter the column names, choose their data types and set the scale and precision. You should also check the Not Null checkbox when you want a column to be mandatory. Click the Next button to continue the create table workflow.

7. After entering the column names, you should choose the data types, enter the scale and precision, and check the NOT NULL checkbox to make appropriate columns mandatory by applying NOT NULL database constraints. If you run out of entry rows, you can click the Add Column button to add new rows. Click the Next button to continue the create table workflow when you’ve defined the columns.

8. After defining the column names, you should choose whether the primary key will use a new sequence or an existing sequence. You also have the ability to not assign a primary key value or simply leave it unpopulated when inserting new rows. The example creates an IMAGE_PK primary key constraint on the IMAGE_ID column, and declares an IMAGE_SEQ sequence value. Click the Next button to continue the create table workflow when you’ve defined the primary key constraint and any new sequence value for the primary key column.

9. After defining the primary key constraint, you can define foreign key column constraints. You enter a foreign key constraint name, choose between a Disallow Delete, Cascade Delete, or Set Null on Delete rule, select the foreign key column, the foreign key’s referenced table and column. Click the Add button to continue the create table workflow.

10. After defining a foreign key constraint, you can see the constraint that you created. Then, you can define another foreign key column constraints. You repeat the steps from the prior steps to add another foreign key constraint. Click the Add button to create a second foreign key constraint and complete the create table workflow.

11. After defining a second foreign key constraint, you see the following two foreign key constraints. Click the Next button to complete the create table workflow.

12. After defining all the foreign key constraints, you can create check and unique constraints. You check a radio button for a check or unique constraint, and then you select the columns for the constraint’s key. Click the /Add button to create any check or unique constraints as part of the create table workflow.

13. After defining all check and unique key constraints, you can see them in the Constraints box. Click the Next button to complete the create table workflow.

14. After defining all items about the table, you can see the SQL to create the IMAGE table and its constraints. You can copy the SQL into a file for later use when writing a re-runnable script. Click the Create button to complete the create table workflow and create the table.

     

    The following are the contents of the script for the actions you’ve defined:

    CREATE table "IMAGE" (
        "IMAGE_ID"         NUMBER NOT NULL,
        "FILE_NAME"        VARCHAR2(60) NOT NULL,
        "MIME_TYPE"        NUMBER NOT NULL,
        "ITEM_IMAGE"       BLOB,
        "CREATED_BY"       NUMBER NOT NULL,
        "CREATION_DATE"    DATE NOT NULL,
        "LAST_UPDATED_BY"  NUMBER NOT NULL,
        "LAST_UPDATE_DATE" DATE NOT NULL,
        constraint  "IMAGE_PK" primary key ("IMAGE_ID")
    )
    /
    
    CREATE sequence "IMAGE_SEQ" 
    /
    
    CREATE trigger "BI_IMAGE"  
      before insert on "IMAGE"              
      for each row 
    begin  
      if :NEW."IMAGE_ID" is null then
        select "IMAGE_SEQ".nextval into :NEW."IMAGE_ID" from dual;
      end if;
    end;
    /   
    
    ALTER TABLE "IMAGE" ADD CONSTRAINT "IMAGE_FK1" 
    FOREIGN KEY ("CREATED_BY")
    REFERENCES "SYSTEM_USER" ("SYSTEM_USER_ID")
    
    /
    ALTER TABLE "IMAGE" ADD CONSTRAINT "IMAGE_FK2" 
    FOREIGN KEY ("LAST_UPDATED_BY")
    REFERENCES "SYSTEM_USER" ("SYSTEM_USER_ID")
    
    /
    alter table "IMAGE" add
    constraint "IMAGE_UK1" 
    unique ("FILE_NAME","MIME_TYPE")
    /   
    

15. After creating the table, trigger, sequence, and constraints, you can see the table definition. You also have the ability to modify the table. At this point, you can create another structure or you can click the Home or SQL Workshop menu choice.

As always, I hope this helps those looking to learn new things and approaches.

The following walks through how you sign on to a STUDENT Workspace with Oracle’s APEX product and write and run free-form SQL statements. You can find instructions on how to create your own STUDENT Workspace.

While this blog introduces several concepts and features of Oracle APEX, it only focuses on how to write and run free-form SQL statements. Overall, Oracle APEX is a valuable tool to learn and master.

1. You start the process by accessing the Oracle Database 11g APEX, which you can access at http://localhost:8080/apex by default on the server. If you’ve got a static IP address for your instance, you can replace localhost with the IP address or hostname for the IP address.
   • Workspace: STUDENT
   • Username:  ADMIN
   • Password:  STUDENT

2. After you login to the STUDENT workspace, you have four options. They are the: Application Builder, SQL Workshop, Team Development, and Administration. You start the process by accessing the Oracle Database 11g APEX, which you can access at http://localhost:8080/apex by default on the server. If you’ve got a static IP address for your instance, you can replace localhost with the IP address or hostname for the IP address. Click on the SQL Workshop icon to proceed.
   • Application Builder: Let’s you build custom APEX applications.
   • SQL Workshop: Let’s you work with custom SQL, and APEX provides you with the following utilities:
     • Object Browser: Lets you create tables, views, and other objects.
     • SQL Commands: Lets you run individual SQL statements inside a browser window and returns results in the bottom pane.
     • SQL Scripts: Lets you create, upload, delete, and run scripts from the browser.
     • Query Builder: Lets you create free form queries that include joins between tables, but limits you to primary to foreign key table relationships. That means you can’t write range joins with a cross join and the BETWEEN operator and you can’t write self-joins.
     • Utilities: Lets you work with the Data Workshop (imports and exports data), Object Reports (a SQL report writer tool), Generate DDL (a tool that creates structures in the database), User Interface Defaults (coordinate data dictionary), Schema Comparison (a tool to compare similarities between schemas, About Database (the ability to connect as the database administrator), and Recycle Bin (dropped and purged structures).
   • Team Development: A project management tool.
   • Administration: Lets you manage database services, users and groups, monitor activities, and dashboards. You should note that the SQL query doesn’t have a semicolon like it would in a SQL*Plus environment. The Run button acts as the execution operator and effectively replaces the role of the semicolon, which traditionally executes a statement.

3. Clicking the SQL Workshop icon takes you to the second level menu. You click the SQL Commands icon to enter a free-form SQL statement. Click on the SQL Commands icon to proceed.

4. The first text panel lets you enter free-form queries. The Autocommit checkbox is enabled, which means the result of INSERT and UPDATE statements are immediate and don’t require a COMMIT statement. The second text panel displays results from a query or acknowledgment of statement completion.

5. This screen shot shows a query in the first panel and the results of the query in the second panel.

As always, I hope this helps those looking to learn new things and approaches.

The question for most new Oracle users is what’s Apex? They have a different question When they discover how to connect to the Oracle Database 11g XE default instance with this URL:

http://localhost:8080/apex

You’ll see the following web site, and wonder what do I enter for the Workspace, the Username, and the Password values?

The answers are:

• Default Workspace: INTERNAL
• Default User: ADMIN
• Default Password: SYS or SYSTEM Password from Install

Enter those values within the initial password time interval and you’ll arrive at the next screen where you can manage the Oracle Database 11g XE instance. If you wait too long, you’ll be redirected to enter the original SYS or SYSTEM password from install and a new password twice. The rules for a new password are:

• Password must contain at least 6 characters.
• New password must differ from old password by at least 2 characters.
• Password must contain at least one numeric character (0123456789).
• Password must contain at least one punctuation character (!”#$%&()“*+,-/:;<=>?_).
• Password must contain at least one upper-case alphabetic character.
• Password must not contain username.

Whether you go directly to the next screen or have to enter your a new password, you should see the following screen:

You can find the default configuration for the installation with the following anonymous PL/SQL block:

DECLARE
  /* Declare variables. */
  lv_endpoint    NUMBER := 1;
  lv_host        VARCHAR2(80);
  lv_port        NUMBER;
  lv_protocol    NUMBER;
BEGIN
  /* Check for current XDB settings. */
  dbms_xdb.getlistenerendpoint( lv_endpoint
                              , lv_host
                              , lv_port
                              , lv_protocol );
 
  /* Print the values. */
  dbms_output.put_line('Endpoint: ['||lv_endpoint||']'||CHR(10)||
                       'Host:     ['||lv_host||']'||CHR(10)||
                       'Port:     ['||lv_port||']'||CHR(10)||
                       'Protocol: ['||lv_protocol||']');
END;
/

It should print the following:

Endpoint: [1]
Host:	  [localhost]
Port:	  [8080]
Protocol: [1]

This is a standalone configuration and you can’t connect to the XDB server from another machine. You can only connect from the local machine.

I hope this helps those trying to use the default Apex 4 installation provided as part of the Oracle Database 11g XE instance. You can read an older post of mine that shows you how to set up a basic Workspace, but after reflection I’ll write more about creating and managing workspaces.

Oracle Database 12c introduces a SEARCH_CONDITION_VC column to the CDB_, DBA_, ALL_, and USER_CONSTRAINTS views. The SEARCH_CONDITION_VC column is a VARCHAR2 data type equivalent to the search condition in the LONG data type SEARCH_CONDITION column. Unfortunately, Oracle Database 11g and earlier versions requires you to convert the LONG data type to a VARCHAR2 for the equivalent behavior. This post provides you with a function to help you do that in Oracle Database 11g.

While Oracle Database 12c let’s you check the search condition of a CHECK constraint, with this query:

SELECT   uc.constraint_name AS constraint_name
,	 uc.search_condition_vc AS search_condition
FROM     user_constraints uc
WHERE    uc.table_name = UPPER('table_name')
AND	 REGEXP_LIKE(uc.search_condition_vc,'search_key','i');

You need the following GET_SEARCH_CONDITION function to convert the SEARCH_CONDITION column from a LONG data type to a VARCHAR2 data type. It uses the DBMS_SQL package to convert the LONG data type.

CREATE OR REPLACE FUNCTION get_search_condition
( pv_table_name   VARCHAR2
, pv_column_name  VARCHAR2 )
RETURN VARCHAR2 AS
 
  /* Declare local variables. */
  lv_cursor           INTEGER := DBMS_SQL.open_cursor;
  lv_feedback         INTEGER;         -- Acknowledgement of dynamic execution
  lv_length           INTEGER;         -- Length of the input string
  lv_value_length     INTEGER;         -- Length of the output string
  lv_constraint_name  VARCHAR2(30);    -- Constraint name
  lv_return           VARCHAR2(32767); -- Function output
  lv_stmt             VARCHAR2(2000);  -- Dynamic SQL statement
  lv_long             LONG;            -- Dynamic LONG data type.
  lv_string           VARCHAR2(32760); -- Maximum length of LONG data type
 
  FUNCTION return_length 
  ( pv_table_name   VARCHAR2
  , pv_column_name  VARCHAR2 ) RETURN VARCHAR2 IS
 
    /* Declare a target variable,  because of the limit of SELECT-INTO. */
    lv_long_view  LONG;
 
    /* Declare a dynamic cursor. */
    CURSOR c
    ( cv_table_name   VARCHAR2
    , cv_column_name  VARCHAR2 ) IS
      SELECT   uc.search_condition
      FROM     user_constraints uc INNER JOIN user_cons_columns ucc
      ON       uc.table_name = ucc.table_name
      AND      uc.constraint_name = ucc.constraint_name
      WHERE    uc.table_name = UPPER(cv_table_name)
      AND      ucc.column_name = UPPER(cv_column_name)
      AND      uc.constraint_type = 'C';
 
  BEGIN
    /* Open, fetch, and close cursor to capture view text. */
    OPEN c (pv_table_name, pv_column_name);
    FETCH c INTO lv_long_view;
    CLOSE c;
 
    /* Return the output CLOB length value. */
    RETURN LENGTH(lv_long_view);
  END return_length;
 
BEGIN
 
  /* Get the length of the CLOB column value. */
  lv_length := return_length(pv_table_name, pv_column_name);
 
  /* Create dynamic statement. */
  lv_stmt := 'SELECT uc.search_condition'||CHR(10)
          || 'FROM   user_constraints uc INNER JOIN user_cons_columns ucc'||CHR(10)
          || 'ON     uc.table_name = ucc.table_name'||CHR(10)
          || 'AND    uc.constraint_name = ucc.constraint_name'||CHR(10)
          || 'WHERE  uc.table_name = UPPER('''||pv_table_name||''')'||CHR(10)
          || 'AND    ucc.column_name = UPPER('''||pv_column_name||''')'||CHR(10)
          || 'AND    uc.constraint_type = ''C''';
 
  /* Parse and define VARCHAR2 and LONG columns. */
  DBMS_SQL.parse(lv_cursor, lv_stmt, DBMS_SQL.native);
  DBMS_SQL.define_column_long(lv_cursor,1);
 
  /* Only attempt to process the return value when fetched. */
  IF DBMS_SQL.execute_and_fetch(lv_cursor) = 1 THEN
    DBMS_SQL.column_value_long(
        lv_cursor
      , 1
      , lv_length
      , 0
      , lv_string
      , lv_value_length);
  END IF;
 
  /* Check for an open cursor. */
  IF DBMS_SQL.is_open(lv_cursor) THEN
    DBMS_SQL.close_cursor(lv_cursor);
  END IF;
 
  /* Convert the long length string to a maximum size length. */
  lv_return := lv_string;
 
  RETURN lv_return;
END get_search_condition;
/

Then, you can use the following query to view the full search criteria of a CHECK constraint that matches part of a search string:

COLUMN constraint_name   FORMAT A16
COLUMN search_condition  FORMAT A30
SELECT   uc.constraint_name AS constraint_name
,        get_search_condition('table_name','column_name') AS search_condition
FROM     user_constraints uc
WHERE    REGEXP_LIKE(get_search_condition('table_name','column_name'),'check_constraint_search_string','i')
AND      uc.constraint_type = 'C';

Hope this helps those looking at discovering the full search criteria of a CHECK constraint.

It seems the Oracle E-Business Suite stymies newbies. Naturally, it’s complex and requires some research before you try to develop any customization.

Having recently fielded a question about how to manage, search, create and update customers for a third party order entry system, I’m always surprised by how little people know about Oracle’s Trading Community Architecture (TCA). They likewise frequently don’t know that there is already a documented process, like the one defined in the Oracle Trading Community Architecture User Guide:

You also have the Oracle Trading Community Architecture Reference Guide, which provides more detailed information. I strongly recommend you read these before trying to Google a solution that may or may not meet your needs. Naturally, good PL/SQL coding skills make this type of activity straightforward. That’s why Oracle staff should master PL/SQL.

The ability to deploy Java inside the Oracle database led somebody to conclude that the source isn’t visible in the data catalog. Then, that person found that they were wrong because the Java source is visible when you use a DDL command to CREATE, REPLACE, and COMPILE the Java source. This post discloses how to find the Java source and how to prevent it from being stored in the data catalog.

You can verify that the Java class and source files exist with the following query:

1
2
3
4
5
6
7
8

COLUMN object_name FORMAT A20 HEADING "Object Name"
COLUMN object_type FORMAT A12 HEADING "Object Type"
COLUMN status      FORMAT A14 HEADING "Object Status"
SELECT   object_name
,        object_type
,        status
FROM     user_objects
WHERE    object_name = 'ReadFile';

It displays:

Object Name          Object Type  Object Status
-------------------- ------------ --------------
ReadFile             JAVA SOURCE  VALID
ReadFile             JAVA CLASS   VALID
 
2 rows selected.

Then, you can use the following query to discovery a Java library created by a SQL command:

1
2
3
4
5
6

COLUMN line FORMAT 9999 HEADING "Line|#"
COLUMN text FORMAT A66  HEADING "Text"
SELECT   line
,        text
FROM     user_source
WHERE    name = 'ReadFile';

It displays the following:

 Line # Text
------- ------------------------------------------------------------------
    1   // Java library imports.
    2   import java.io.File;
    3   import java.io.BufferedReader;
    4   import java.io.FileNotFoundException;
    5   import java.io.IOException;
    6   import java.io.FileReader;
    7   import java.security.AccessControlException;
    8
    9   // Class definition.
   10   public class ReadFile {
   11     // Define class variables.
   12     private static File file;
   13     private static FileReader inTextFile;
   14     private static BufferedReader inTextReader;
   15     private static StringBuffer output = new StringBuffer();
   16     private static String outLine, outText;
   17
   18     // Define readText() method.
   19     public static String readText(String fromFile)
   20       throws AccessControlException, IOException {
   21       // Read file.
   22       try {
   23         // Initialize File.
   24         file = new File(fromFile);
   25
   26         // Check for valid file.
   27         if (file.exists()) {
   28
   29           // Assign file to a stream.
   30           inTextFile = new FileReader(file);
   31           inTextReader = new BufferedReader(inTextFile);
   32
   33           // Read character-by-character.
   34           while ((outLine = inTextReader.readLine()) != null) {
   35             output.append(outLine + "\n"); }
   36
   37           // Assing the StringBuffer to a String.
   38           outText = Integer.toString(output.toString().length());
   39
   40           // Close File.
   41           inTextFile.close(); }
   42         else {
   43           outText = new String("Empty"); }}
   44       catch (IOException e) {
   45         outText = new String("");
   46         return outText; }
   47     return outText; }}
 
47 rows selected.

You can eliminate the source by compiling the Java library outside the database. Then, you use the loadjava utility to load the only the class file into the data catalog. The syntax would be the following command for an importer user in a video Pluggable Database (PDB):

loadjava -r -f -o -user importer/importer@video ReadFile.class

You should know that this syntax is disallowed by the loadjava utility, notwithstanding it’s found in the Oracle Database 12c documentation:

loadjava -r -f -o -user importer@video/importer ReadFile.class

You can verify that only the Java class file exists with the following query:

1
2
3
4
5
6
7
8

COLUMN object_name FORMAT A20 HEADING "Object Name"
COLUMN object_type FORMAT A12 HEADING "Object Type"
COLUMN status      FORMAT A14 HEADING "Object Status"
SELECT   object_name
,        object_type
,        status
FROM     user_objects
WHERE    object_name = 'ReadFile';

It displays:

Object Name          Object Type  Object Status
-------------------- ------------ --------------
ReadFile             JAVA CLASS   VALID
 
1 row selected.

Hope this helps those who want to hide the Java source files.

A few years ago, I gave a session on Oracle’s Data Vault. An attendee from that session and I happened to be seated at the same table for lunch last Thursday when I presented at Utah Oracle User Group Fall Symposium. He asked if I’d ever found a way to encrypt a column from the prying eyes of the SYS user. I said yes, and he asked how. It was a bit more than could be explained at lunch and promised to put it on the blog. (Disclaimer: It is possible to unwrap wrapped code but the mechanics change with each release and they require access to the SYS schema and substantial Application DBA knowledge; however, there are scripts published on the Internet to unwrap the code. There are also other vulnerabilities in the example, but I’ll pass on exposing them. After all this was written to illustrate an approach.)

It’s important to note you can now hide columns in Oracle Database 12c, but they’re still visible to the DBA-level staff. That’s why I’ll publish a new article on re-writing this encrypted object as a Java library, and disclose how to hide password exchanges from the SGA area.

This demonstrates how you can encapsulate a column from even the SYS user. I also put it into Appendix D, PL/SQL Built-in Packages and Types to illustrate the DBMS_CRYPTO package. It uses a User Defined Type (UDT) and assumes you have a working knowledge of object types in Oracle 10g forward. If not, you can find them in:

1. Chapter 11 of my Oracle Database 12c PL/SQL Programming book.
2. Chapter 14 of my Oracle Database 11g PL/SQL Programming book.

It also assumes basic knowledge of Oracle’s encryption technology, some of which is explained in the new Oracle Database 12c PL/SQL Programming book.

For reference, a good DBA would simply find this clear text password in the SGA. A more secure approach might be hiding the encryption keyword in a one column and table or embedded in some string within a common lookup table column value as a position specific substring. Alas, I don’t have time to write something so elaborate.

The steps are:

1. You must create a user defined type (UDT), which sets the up a single salary column.

1
2
3
4
5
6
7
8
9
10

CREATE OR REPLACE TYPE masked IS OBJECT
( salary  RAW(1000)
, CONSTRUCTOR FUNCTION masked RETURN SELF AS RESULT
, CONSTRUCTOR FUNCTION masked ( salary  NUMBER ) RETURN SELF AS RESULT
, MEMBER FUNCTION get_raw_salary RETURN RAW
, MEMBER FUNCTION get_salary ( KEY VARCHAR2 ) RETURN NUMBER
, MEMBER PROCEDURE to_string
, ORDER MEMBER FUNCTION equals ( object MASKED ) RETURN NUMBER )
INSTANTIABLE FINAL;
/

2. As the SYS user, you must grant EXECUTE privilege on the DBMS_CRYPTO package to the target SCHEMA user.

1

GRANT EXECUTE ON dbms_crypto TO schema_name;

3. You implement the MASKED UDT, as follows below (source unencrypted, don’t worry I show you how to encrypt [wrap] it in a moment).

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105

CREATE OR REPLACE TYPE BODY masked IS
 
  CONSTRUCTOR FUNCTION masked RETURN SELF AS RESULT IS
 
    /* Create a placeholder for a zero salary, like an empty_clob() call. */
    zero MASKED := masked(0);
 
  BEGIN
 
    /* Assign an encrypted zero salary to the instance. */
    self := zero;  
    RETURN;
 
  END masked;
  CONSTRUCTOR FUNCTION masked ( salary NUMBER ) RETURN SELF AS RESULT IS
 
    /* Declare local variables for encryption, object types hold instance
       objects and object body variables are method specific. Placing them
       inside the methods while tedious prevents their disclosure. */
    lv_key_string      VARCHAR2(4000)  := 'Encrypt Me!'; 
    lv_key             RAW(1000);
    lv_raw             RAW(1000);
    lv_encrypted_data  RAW(1000);
 
  BEGIN
    /* Dynamic assignment. */
    lv_raw := UTL_RAW.cast_to_raw(NVL(salary,0));
 
    /* Convert to a RAW 64-character key. */
    lv_key := UTL_RAW.cast_to_raw(lv_key_string);
    lv_key := RPAD(lv_key,64,'0');   
 
    /* Encrypt the salary before assigning it to the object type attribute */
    lv_encrypted_data := DBMS_CRYPTO.ENCRYPT(lv_raw, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5,lv_key);
    self.salary := lv_encrypted_data;
 
    RETURN;
  END masked;
 
  MEMBER FUNCTION get_raw_salary RETURN RAW IS 
  BEGIN
    RETURN self.salary;
  END get_raw_salary;    
 
  MEMBER FUNCTION get_salary( key VARCHAR2 ) RETURN NUMBER IS 
 
    /* Declare local variables for encryption, object types hold instance
       objects and object body variables are method specific. Placing them
       inside the methods while tedious prevents their disclosure. */
    lv_key_string      VARCHAR2(4000)  := 'Encrypt Me!'; 
    lv_decrypted_data  RAW(4000);
    lv_key             RAW(1000);
    lv_return_value    NUMBER;
 
  BEGIN
 
    /* Verify key value matches local value before decrypting, substitute
       a zero value when the key doesn't match. */
    IF key = lv_key_string THEN
      lv_key := UTL_RAW.cast_to_raw(lv_key_string);
      lv_key := RPAD(lv_key,64,'0');   
      lv_decrypted_data := DBMS_CRYPTO.DECRYPT(self.salary, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5,lv_key);
      lv_return_value := TO_NUMBER(TO_CHAR(UTL_RAW.cast_to_number(lv_decrypted_data),'9999990.00'));
    ELSE
      lv_return_value := 0;    
    END IF;
 
    RETURN lv_return_value;
  END get_salary;
 
  ORDER MEMBER FUNCTION equals ( object MASKED ) RETURN NUMBER IS
 
    /* Declare local variables for encryption, object types hold instance
       objects and object body variables are method specific. Placing them
       inside the methods while tedious prevents their disclosure. */
    lv_key_string      VARCHAR2(4000)  := 'Encrypt Me!'; 
    lv_decrypted_self  RAW(4000);
    lv_decrypted_peer  RAW(4000);
    lv_key             RAW(1000);
 
  BEGIN
 
    /* Decrypt the current and peer object attribute values before
       comparing their values. */  
    lv_key := UTL_RAW.cast_to_raw(lv_key_string);
    lv_key := RPAD(lv_key,64,'0');   
    lv_decrypted_self := DBMS_CRYPTO.DECRYPT(self.salary, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5, lv_key);
    lv_decrypted_peer := DBMS_CRYPTO.DECRYPT(object.get_raw_salary(), dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5, lv_key);
 
    /* Sort order key. */
    IF lv_decrypted_self < lv_decrypted_peer THEN
      RETURN 1;
    ELSE
      RETURN 0;
    END IF;
 
  END equals;
 
  MEMBER PROCEDURE to_string IS
  BEGIN
    /* Provide a to_string method for good practice. */
    DBMS_OUTPUT.put_line('Encrypted value');
  END to_string;
END;
/

4. You implement the MASKED UDT encrypted by using the DBMS_DDL package, as follows below. This ensures that others can’t read the source code by querying the ALL_, DBA_, or USER_SOURCE views. You should note that I’ve removed comments and unnecessary spaces.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64

BEGIN
  DBMS_DDL.create_wrapped(
     'CREATE OR REPLACE TYPE BODY masked IS '
  || 'CONSTRUCTOR FUNCTION masked RETURN SELF AS RESULT IS '
  || 'zero MASKED := masked(0); '
  || 'BEGIN '
  || 'self := zero; '
  || 'RETURN; '
  || 'END masked; '
  || 'CONSTRUCTOR FUNCTION masked ( salary NUMBER ) RETURN SELF AS RESULT IS '
  || 'lv_key_string      VARCHAR2(4000)  := ''Encrypt Me!''; '
  || 'lv_key             RAW(1000); '
  || 'lv_raw             RAW(1000) := RPAD(utl_raw.cast_from_number(salary),32,''0''); '
  || 'lv_encrypted_data  RAW (1000); '
  || 'BEGIN '
  || 'lv_key := utl_raw.cast_to_raw(lv_key_string); '
  || 'lv_key := RPAD(lv_key,64,''0''); '
  || 'lv_encrypted_data := DBMS_CRYPTO.ENCRYPT(lv_raw, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5,lv_key); '
  || 'self.salary := lv_encrypted_data; '
  || 'RETURN; '
  || 'END masked; '
  || 'MEMBER FUNCTION get_raw_salary RETURN RAW IS '
  || 'BEGIN '
  || 'RETURN self.salary; '
  || 'END get_raw_salary; '
  || 'MEMBER FUNCTION get_salary( key VARCHAR2 ) RETURN NUMBER IS '
  || 'lv_key_string      VARCHAR2(4000)  := ''Encrypt Me!''; '
  || 'lv_decrypted_data  RAW(4000); '
  || 'lv_key             RAW(1000); '
  || 'lv_return_value    NUMBER; '
  || 'BEGIN '
  || 'IF key = lv_key_string THEN '
  || 'lv_key := utl_raw.cast_to_raw(lv_key_string); '
  || 'lv_key := RPAD(lv_key,64,''0''); '
  || 'lv_decrypted_data := DBMS_CRYPTO.DECRYPT(self.salary, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5,lv_key); '
  || 'lv_return_value := TO_NUMBER(TO_CHAR(utl_raw.cast_to_number(lv_decrypted_data),''9999990.00'')); '
  || 'ELSE '
  || 'lv_return_value := 0; '
  || 'END IF; '
  || 'RETURN lv_return_value; '
  || 'END get_salary; '
  || 'ORDER MEMBER FUNCTION equals ( object MASKED ) RETURN NUMBER IS '
  || 'lv_key_string      VARCHAR2(4000)  := ''Encrypt Me!''; '
  || 'lv_decrypted_self  RAW(4000); '
  || 'lv_decrypted_peer  RAW(4000); '
  || 'lv_key             RAW(1000); '
  || 'BEGIN '
  || 'lv_key := utl_raw.cast_to_raw(lv_key_string);'
  || 'lv_key := RPAD(lv_key,64,''0''); '
  || 'lv_decrypted_self := DBMS_CRYPTO.DECRYPT(self.salary, dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5, lv_key); '
  || 'lv_decrypted_peer := DBMS_CRYPTO.DECRYPT(object.get_raw_salary(), dbms_crypto.ENCRYPT_AES256 + dbms_crypto.CHAIN_CBC + dbms_crypto.PAD_PKCS5, lv_key); '
  || 'IF lv_decrypted_self < lv_decrypted_peer THEN '
  || 'RETURN 1; '
  || 'ELSE '
  || 'RETURN 0; '
  || 'END IF; '
  || 'END equals; '
  || 'MEMBER PROCEDURE to_string IS '
  || 'BEGIN '
  || 'dbms_output.put_line(''Encrypted value''); '
  || 'END to_string; '
  || 'END; ');
END;
/

5. You can test a single instance with this anonymous PL/SQL block.

1
2
3
4
5
6
7
8
9

DECLARE
  o MASKED := masked(82000.12);
BEGIN
  DBMS_OUTPUT.put('Override:  ');
  o.to_string();
  DBMS_OUTPUT.put_line('Decrypted: '||o.get_salary('Encrypt Me!'));
  DBMS_OUTPUT.put_line('Bad Key:   '||o.get_salary('Incorrect'));
END;
/

Override:  Encrypted value
Decrypted: 82000.12
Bad Key:   0

6. You can test a series of instances by making them persistent objects, or columns in a table, and then query values from them. It also lets you you test the sorting feature provided in the UDT.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

CREATE TABLE sort_demo (salary MASKED);
 
INSERT INTO sort_demo VALUES (masked(82000.24));
INSERT INTO sort_demo VALUES (masked(61000.12));
INSERT INTO sort_demo VALUES (masked(93000.36));
 
SELECT salary AS "Encrypted" FROM sort_demo;
 
COLUMN unordered FORMAT 9,999,990.00 HEADING "Unordered|List"
 
SELECT TREAT(salary AS MASKED).get_salary('Encrypt Me!') AS Unordered
FROM   sort_demo;
 
COLUMN ordered FORMAT 9,999,990.00 HEADING "Ordered|List"
 
SELECT TREAT(salary AS MASKED).get_salary('Encrypt Me!') AS Ordered
FROM   sort_demo
ORDER BY 1;

Encrypted(SALARY)
--------------------------------------------------------------------------
MASKED('1798C04F8342C53A826144568075CBDB145D0C9BD226B410C8C7167B76382D86')
MASKED('82F783F2E117AA60955B0A2E73545506936D6F8FFBEC5D9E0D8E70B82D4B694D')
MASKED('1551F350AAEB30ADFC1527F25CAA935732243858AF1C5D724A78B997A4394EAD')
 
    Unordered
         List
-------------
    82,000.24
    61,000.12
    93,000.36
 
 
      Ordered
         List
-------------
    61,000.12
    82,000.24
    93,000.36

If you want to retest this, make sure you drop the SORT_DEMO table first. Hope this helps any interested in an elaborate solution.

After the installation of Oracle 11g XE on Windows, I anticipated problems installing on Fedora. It’s terrific to report it was by far simpler to install on Fedora. This post provides the steps, and builds on the Fedora configuration published earlier this week.

1. After you download the software from the Oracle web site, you’ll need to expand the compressed file. When you double click on the download item you will see the following screen. Click Extract at the top of the display.

2. The Extract displays the following dialog. Click the Create Folder button and you’ll get an entry point for a new directory in your user’s directory. For the example, Oracle11gXE

3. After creating the directory, click the Extract button.

4. The extract process completes and shows the following dialog.

5. The Disk1 directory will show the following contents.

6. You need to drop down into a Terminal session, which you can launch by clicking on Applications, System Tools, and Terminal in the menu.

sudo rpm -iv oracle-xe-11.2.0-1.0.x86_64.rpm

[sudo] password for mclaughlinm:
Preparing packages for installation...
oracle-xe-11.2.0-1.0
Executing post-install steps...
You must run '/etc/init.d/oracle-xe configure' as the root user to configure the database.

7. This step requires that you assume the role of the root user, which can be done with this syntax:

sudo sh

/etc/init.d/oracle-xe configure

Oracle Database 11g Express Edition Configuration
-------------------------------------------------
This will configure on-boot properties of Oracle Database 11g Express 
Edition.  The following questions will determine whether the database should 
be starting upon system boot, the ports it will use, and the passwords that 
will be used for database accounts.  Press <Enter> to accept the defaults. 
Ctrl-C will abort.
 
Specify the HTTP port that will be used for Oracle Application Express [8080]:
 
Specify a port that will be used for the database listener [1521]:
 
Specify a password to be used for database accounts.  Note that the same
password will be used for SYS and SYSTEM.  Oracle recommends the use of 
different passwords for each database account.  This can be done after 
initial configuration:
Confirm the password:
 
Do you want Oracle Database 11g Express Edition to be started on boot (y/n) [y]:
 
Starting Oracle Net Listener...Done
Configuring database...
Starting Oracle Database 11g Express Edition instance...Done
Installation completed successfully.

8. This step requires that you exit the root shell by typing the exit command. This should put you back into your administration account with sudoer privileges. You use vi to edit and add the following to the .bashrc file.

. /u01/app/oracle/product/11.2.0/xe/bin/oracle_env.sh

# Oracle Settings
TMP=/tmp; export TMP
TMPDIR=$TMP; export TMPDIR
 
ORACLE_HOSTNAME=localhost.localdomain; export ORACLE_HOSTNAME
ORACLE_UNQNAME=DB11G; export ORACLE_UNQNAME
ORACLE_BASE=/u01/app/oracle; export ORACLE_BASE
ORACLE_HOME=$ORACLE_BASE/product/11.2.0/xe; export ORACLE_HOME
ORACLE_SID=XE; export ORACLE_SID
NLS_LANG=`$ORACLE_HOME/bin/nls_lang.sh`; export NLS_LANG
ORACLE_TERM=xterm; export ORACLE_TERM
PATH=/usr/sbin:$PATH; export PATH
PATH=$ORACLE_HOME/bin:$PATH; export PATH
 
LD_LIBRARY_PATH=$ORACLE_HOME/lib:/lib:/usr/lib; export LD_LIBRARY_PATH
CLASSPATH=$ORACLE_HOME/JRE:$ORACLE_HOME/jlib:$ORACLE_HOME/rdbms/jlib; export CLASSPATH
 
if [ $USER = "oracle" ]; then
  if [ $SHELL = "/bin/ksh" ]; then
    ulimit -p 16384
    ulimit -n 65536
  else
    ulimit -u 16384 -n 65536
  fi
fi

sqlplus system

9. This step is optional. You can configure an oracleenv.sh file in the oracle user’s home directory. While you could put this in the .bashrc file, the oracle account isn’t bootable. That’s why I recommend putting it in an environment file.

# .bashrc
 
# Source global definitions
if [ -f /etc/bashrc ]; then
  . /etc/bashrc
fi
 
# Uncomment the following line if you don't like systemctl's auto-paging feature:
# export SYSTEMD_PAGER=
 
# User specific aliases and functions
. /u01/app/oracle/product/11.2.0/xe/bin/oracle_env.sh
 
# Wrap sqlplus with rlwrap to edit prior lines with the
# up, down, left and right keys.
sqlplus()
{
  if [ "$RLWRAP" = "0" ]; then
    sqlplus "$@"
  else
    rlwrap sqlplus "$@"
  fi
}
 
# Set vi as a command line editor.
set -o vi

If you want to use the menu choices added by the installation, you need to put your privileged sudoer in the dba group. I wrote a blog post on it here.