MacLochlainns Weblog

Have you ever wondered how to leverage substitution variables in anonymous block programs? There are several tricks that you can use beyond passing numeric and string values to local variable. The generic default appears to take a number unless you cast it as a string but that’s not really the whole story. The first two are standard examples of how to use numeric and string substitution values.

The following accept a numeric substitution value:

1
2
3
4
5
6
7
8
9
10

DECLARE
  lv_input  NUMBER;
BEGIN
  /* Assign substitution value to local variable. */
  lv_input := &input;
 
  /* Print the local variable. */
  dbms_output.put_line('['||lv_input||']');
END;
/

The following accept a string substitution value, casts the input as a string, assigns the string value to a 4,000 character length local variable, checks whether the 4,000 character length is greater than 10, and assigns the first 10 characters to the lv_parse_input variable:

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

DECLARE
  lv_unparsed_input  VARCHAR2(4000);
  lv_parsed_input    VARCHAR2(10);
BEGIN
  /* Assign substitution value to local variable. */
  lv_unparsed_input := '&input';
 
  /* Check size of input value. */
  IF LENGTH(lv_unparsed_input) > 10 THEN
    lv_parsed_input := SUBSTR(lv_unparsed_input,1,10);
  END IF;
 
  /* Print the local variable. */
  dbms_output.put_line('Print {lv_parsed_input}: ['||lv_parsed_input||']');
END;
/

Next, let’s examine two tricks. The first passes a case insensitive variable name and the second passes a case sensitive variable name as a parameter to an anonymous block program.

This declares an anonymous block program that uses a substitution value as a variable name:

1
2
3
4
5
6
7

DECLARE
  mine  VARCHAR2(10) := 'Default';
BEGIN
  /* Print the local variable's value. */
  dbms_output.put_line('Print {mine} variable value: ['||&input||']');
END;
/

When you run the anonymous block, you’re prompted for an input variable. You provide a case insensitive variable name as the input value:

Enter value for input: MINE
old   5:   dbms_output.put_line('['||&input||']');
new   5:   dbms_output.put_line('['||MINE||']');
Print {mine} variable value: [Default]

The downside of this approach, yields an ORA-06550 and PLS-00201 exception. Neither of these can be caught because Oracle raises the errors during parsing when the variable name isn’t a 100% case insensitive match. The same type of problem occurs in the next example when the input variable isn’t a 100% case sensitive match.

You can rewrite the program to handle case insensitive variables like this:

1
2
3
4
5
6
7

DECLARE
  "Mine"  VARCHAR2(10) := 'Default';
BEGIN
  /* Print the local variable's value. */
  dbms_output.put_line('Print {mine} variable value: ['||"&input"||']');
END;
/

When you run the anonymous block, you’re prompted for an input variable. You provide a case sensitive variable name as the input value:

Enter value for input: Mine
old   5:   dbms_output.put_line('['||&input||']');
new   5:   dbms_output.put_line('['||"Mine"||']');
Print {Mine} variable value: [Default]

Hope this helps those looking for a solution.

Using Oracle’s Resource Manager requires you to understand the IO dynamics. The first step requires you to run the CALIBRATE_IO procedure from the DBMS_RESOURCE_MANAGER package.

Oracle provides some great examples about how to use the CALIBRATE_IO procedure of the DBMS_RESOURCE_MANAGER package in the Oracle Database Database PL/SQL Packages and Types Reference. The CALIBRATE_IO procedure returns the best answer when you provide a valid number of files, which you can capture by querying the V$ASM_DISK view.

The following code queries the view and assigns the value to a session level variable:

CLEAR BREAKS
CLEAR COLUMNS
CLEAR COMPUTES
 
VARIABLE files NUMBER
 
BEGIN
  SELECT COUNT(DISTINCT name) disks
  INTO :files
  FROM v$asm_disk;
END;
/

When you have the number of files, you can calibrate the IO with the following anonymous block. The query should always work but just in case the NVL function on line 9 assigns the default number of files.

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

DECLARE
  lv_num_physical_disks BINARY_INTEGER; — v$asm_disk
  lv_max_latency BINARY_INTEGER := 10;
  lv_max_iops BINARY_INTEGER;
  lv_max_mbps BINARY_INTEGER;
  lv_actual_latency BINARY_INTEGER;
BEGIN
  /* Assign actual files to anonymous block variable. */
  lv_num_physical_disks := NVL(:files,2);
 
  /* Run the calibrate_io procedure. */
  DBMS_RESOURCE_MANAGER.CALIBRATE_IO(
      num_physical_disks => lv_num_physical_disks
    , max_latency => lv_max_latency
    , max_iops => lv_max_iops
    , max_mbps => lv_max_mbps
    , actual_latency => lv_actual_latency);
END;
/

You can query the results like this:

SELECT max_iops
,      max_mbps
,      max_pmbps
,      latency
,      num_physical_disks
FROM   dba_rsrc_io_calibrate;

It should show results like these:

MAX_IOPS MAX_MBPS MAX_PMBPS LATENCY NUM_PHYSICAL_DISKS
-------- -------- --------- ------- ------------------
    8894      443       294       9                 18

Hope this helps those using the CALIBRATE_IO procedure of the DBMS_RESOURCE_MANAGER package.

Who can resist a free Rich Nimeiec book on SQL Tuning? O.K., those who know everything can resist. If you’re like me, this is an opportunity to learn from Rich. Click on the book image or this link to get a free copy, or if you want to pay $10 for a copy click here to buy Quick Start Guide to Oracle Query Tuning: Tips for DBAs and Developers from Amazon.com.

The book is four chapters long, is a 129 pages in length, and is in a PDF format. The outline is:

1. Query Tuning: Developer and Beginning DBA
2. Query Tuning: Basics for DBAs and Developers
3. Advanced Performance Tuning
4. Tips for Tuning When You Have Everything Tuned

Enjoy reading it. His more comprehensive book is Oracle Database 11g Release 2 Performance Tuning Tips & Techniques (Oracle Press) and it’s $30, but it’s written for an advanced audience (more or less OCA or higher).

Using an Oracle object type’s instance in a query is a powerful capability. Unfortunately, Oracle’s SQL syntax doesn’t make it immediately obvious how to do it. Most get far enough to put it in a runtime view (a subquery in the FROM clause), but then they get errors like this:

SELECT	 instance.get_type()
         *
ERROR AT line 4:
ORA-00904: "INSTANCE"."GET_TYPE": invalid identifier

The problem is how Oracle treats runtime views, which appears to me as a casting error. Somewhat like the ORDER BY clause irregularity that I noted in July, the trick is complete versus incomplete syntax. The following query fails and generates the foregoing error:

1
2
3
4

SELECT instance.get_type() AS object_type
,      instance.to_string() AS object_content
FROM  (SELECT dependent()AS instance
       FROM   dual);

If you add a table alias, or name, to the runtime view on line 4, it works fine:

1
2
3
4

SELECT cte.instance.get_type() AS object_type
,      cte.instance.to_string() AS object_content
FROM  (SELECT dependent() AS instance
       FROM   dual) cte;

That is the trick. You use an alias for the query, which assigns the alias like a table reference. The reference lets you access instance methods in the scope of a query. Different columns in the query’s SELECT-list may return different results from different methods from the same instance of the object type.

You can also raise an exception if you forget the open and close parentheses for a method call to a UDT, which differs from how Oracle treats no argument functions and procedures. That type of error would look like this:

SELECT cte.instance.get_type AS object_type
       *
ERROR AT line 1:
ORA-00904: : invalid identifier

It is an invalid identifier because there’s no public variable get_type, and a method is only found by using the parenthesis and a list of parameters where they’re required.

The object source code is visible by clicking on the expandable label below.

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

DROP TYPE dependent_t;
DROP TYPE base_t;
 
SET SERVEROUTPUT ON SIZE UNLIMITED
 
-- Create an object.
CREATE OR REPLACE TYPE base_t IS OBJECT
( TYPE       VARCHAR2(20)
, CONSTRUCTOR FUNCTION base_t RETURN SELF AS RESULT
, MEMBER FUNCTION get_type RETURN VARCHAR2
, MEMBER FUNCTION to_string RETURN VARCHAR2) 
INSTANTIABLE NOT FINAL;
/
 
-- Create an object body.
CREATE OR REPLACE TYPE BODY base_t IS
CONSTRUCTOR FUNCTION base_t RETURN SELF AS RESULT IS
BEGIN
  RETURN;
END base_t;
MEMBER FUNCTION get_type RETURN VARCHAR2 IS
BEGIN
  RETURN self.TYPE;
END;
MEMBER FUNCTION to_string RETURN VARCHAR2 IS
BEGIN
  RETURN self.TYPE;
END to_string;
END;
/
 
-- Create a subtype.
CREATE OR REPLACE TYPE dependent UNDER base_t
( child  VARCHAR2(40)
, CONSTRUCTOR FUNCTION dependent RETURN SELF AS RESULT
, CONSTRUCTOR FUNCTION dependent
  ( child  VARCHAR2 ) RETURN SELF AS RESULT
, OVERRIDING MEMBER FUNCTION get_type RETURN VARCHAR2
, OVERRIDING MEMBER FUNCTION to_string RETURN VARCHAR2) 
INSTANTIABLE NOT FINAL;
/
 
-- Create a subtype body.
CREATE OR REPLACE TYPE BODY dependent IS
  CONSTRUCTOR FUNCTION dependent RETURN SELF AS RESULT IS
  BEGIN
    /* Assign subtype name to type. */
    self.TYPE := 'DEPENDENT';
    RETURN;
  END dependent;
  CONSTRUCTOR FUNCTION dependent
  ( child  VARCHAR2 ) RETURN SELF AS RESULT IS
    lv_dependent DEPENDENT := dependent();
  BEGIN
    /* Assign default constructor to self instance. */
    self := lv_dependent;
    /* Assign parameters to object instance. */
    self.TYPE  := 'DEPENDENT';
    self.child := child;
    RETURN;
  END dependent;
  OVERRIDING MEMBER FUNCTION get_type RETURN VARCHAR2 IS
  BEGIN
    RETURN self.TYPE;
  END;
  OVERRIDING MEMBER FUNCTION to_string RETURN VARCHAR2 IS
  BEGIN
    RETURN self.TYPE;
  END to_string;
END;
/

As always, I hope this helps those solving problems.

Sometimes I give parts of a solution to increase the play time to solve a problem. I didn’t anticipate a problem when showing how to perform a sort operation with a CASE statement. It’s a sweet solution when you need to sort something differently than a traditional ascending or descending sort.

I gave my students this ORDER BY clause as an example:

  CASE
    WHEN filter = 'Debit' THEN 1
    WHEN filter = 'Credit' THEN 2
    WHEN filter = 'Total' THEN 3
  END;

It raises the following error in MySQL for students:

ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'ORDER BY
  CASE
    WHEN filter = 'Debit' THEN 1
    WHEN filter = 'Credit' THEN' at line 6

It raises the following error in Oracle for some students:

  CASE
  *
ERROR AT line 7:
ORA-01785: ORDER BY item must be the NUMBER OF a SELECT-list expression

So, I built a little test case to replicate the problem and error message they encountered:

SQL> SELECT 'Debit' AS filter FROM dual
  2  UNION ALL
  3  SELECT 'Credit' AS filter FROM dual
  4  UNION ALL
  5  SELECT 'Total' AS filter FROM dual
  6  ORDER BY
  7    CASE
  8      WHEN filter = 'Debit' THEN 1
  9      WHEN filter = 'Credit' THEN 2
 10      WHEN filter = 'Total' THEN 3
 11    END;

They said, great but how can you fix it? That’s simple, with a Common Table Expression (CTE) in Oracle or with an inline view in MySQL. The Oracle CTE solution is:

  1  WITH results AS
  2  (SELECT 'Debit' AS filter FROM dual
  3   UNION ALL
  4   SELECT 'Credit' AS filter FROM dual
  5   UNION ALL
  6   SELECT 'Total' AS filter FROM dual)
  7  SELECT filter
  8  FROM   results
  9  ORDER BY
 10    CASE
 11	 WHEN filter = 'Debit'  THEN 1
 12	 WHEN filter = 'Credit' THEN 2
 13	 WHEN filter = 'Total'  THEN 3
 14    END;

There are two MySQL solutions. One simply removes the FROM dual clauses from the query components and the other uses an inline view in the FROM clause. This is the inline view:

SELECT filter
FROM  (SELECT 'Debit' AS filter FROM dual
       UNION ALL
       SELECT 'Credit' AS filter FROM dual
       UNION ALL
       SELECT 'Total' AS filter FROM dual) resultset
ORDER BY
  CASE
    WHEN filter = 'Debit' THEN 1
    WHEN filter = 'Credit' THEN 2
    WHEN filter = 'Total' THEN 3
  END;

This is the solution without the FROM dual clauses:

SELECT 'Debit' AS filter
UNION ALL
SELECT 'Credit' AS filter
UNION ALL
SELECT 'Total' AS filter
ORDER BY
  CASE
    WHEN filter = 'Debit' THEN 1
    WHEN filter = 'Credit' THEN 2
    WHEN filter = 'Total' THEN 3
  END;

Both MySQL solutions yield the following:

+--------+
| filter |
+--------+
| Debit  |
| Credit |
| Total  |
+--------+
3 rows in set (0.00 sec)

It puts the fabricating query inside a result set, and then lets you use the column alias to filter the set. If you have a better approach, please share it.

This how you install SQL Developer on Mac OS Yosemite. The first thing you need to do is download and install Java 8, not Java 7 on your Mac OS Yosemite as suggested on some web sites. You can determine whether or not Java is installed by running the following command:

Mac-Pro-3:~ username$ java -version
No Java runtime present, requesting install.

You must accept the Java license to install Java 8 on the Mac OS X operating system:

You have the option of installing the Java SDK or JDK. I’ve opted to install Netbeans 8 with JDK 8u45, as you can tell from the screen capture after you launched the file:

It is a standard Mac OS installation, which is why I didn’t bother showing any dialog messages. After installing the Java JDK or SDK, you should download SQL Developer 4.1 from Oracle’s web site. Below is a screen shot of the Oracle download web page where I’ve accepted the license agreement:

If you attempt to launch the installation and you’ve set your Mac Security to the “Mac App Store and identified developers” setting, you should raise the following exception:

If you reset the Mac Security to an “Anywhere” setting, you can install Oracle SQL Developer on Yosemite. Just make sure you reset it to the “Mac App Store and identified developers” setting after you install SQL Developer.

If you launch SQL Developer with the Security “Anywhere” setting, it displays the following dialog:

After you launch the program, you will see the following progress dialog:

The last step of the installation launches SQL Developer, as shown below:

Click the Connections icon to create an initial connection, like the following:

After connecting to the database, you can write and execute a query as shown in the next screen capture:

As always, I hope that this helps those who require an example to install SQL Server on a Mac OS.

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.

A reader posted A dynamic level limiting hierarchical query about Oracle’s hierarchical queries. They wanted to know how to capture only the hierarchy above the level where the first leaf node occurs. They gave me the following hierarchy map as an example:

               1                                    2
        +-------------+                       +-----------+
        |             |                       |           |      
        3             5                       4           6
    +---------+    +-----------+           +-----+    +------+
    |         |    |           |           |     |    |      |
    7         9    11          13          8     10   12     14
+-----+   +-----+  +--+    +-------+                       +-----+ 
|     |   |     |     |    |       |                       |     |
15    17  19    21    23   27      29                     16     18
                                                                 +---+
                                                                     |
                                                                     20

You can find the node values and hierarchical level with the following query:

SELECT   tt.child_id
,        LEVEL
FROM     test_temp tt
WHERE    CONNECT_BY_ISLEAF  = 1
START
WITH     tt.parent_id IS NULL 
CONNECT
BY PRIOR tt.child_id = tt.parent_id
ORDER BY 2;

We really don’t need the node values to solve the problem. We only need the lowest LEVEL value returned by the query, which is 3. The combination of the MIN and CONNECT_BY_ISLEAF functions let us solve this problem without writing a PL/SQL solution. The subquery returns the lowest level value, which is the first level where a leaf node occurs.

SELECT   LPAD(' ', 2*(LEVEL - 1)) || tt.child_id AS child_id
FROM     test_temp tt
WHERE    LEVEL <= (SELECT   MIN(LEVEL)
                   FROM     test_temp tt
                   WHERE    CONNECT_BY_ISLEAF  = 1
                   START
                   WITH     tt.parent_id IS NULL 
                   CONNECT
                   BY PRIOR tt.child_id = tt.parent_id)
START
WITH     tt.parent_id IS NULL
CONNECT
BY PRIOR tt.child_id = tt.parent_id;

It returns:

               1                                    2
        +-------------+                       +-----------+
        |             |                       |           |      
        3             5                       4           6
    +---------+    +-----------+           +-----+    +------+
    |         |    |           |           |     |    |      |
    7         9    11          13          8     10   12     14

While I answered the question in a comment originally, it seemed an important trick that should be shared in its own post.

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.

Planning out my year, I decided to take the Oracle OCP and MySQL OCP exams. I checked for review books and was pleasantly surprised to find the soon to be released OCP MySQL Database Administrator Exam Guide (Exam 1Z0-883). However, I noticed that the book was actually prepared for the obsolete and discountinued Exams 1Z0-870, 1Z0-873, and 1Z0-874. As it turns out, Steve O’Hearn has informed me that there isn’t a book and that the posting in Amazon.com is in error.

There isn’t an alternative review book for the OCP MySQL 5.6 Developer or Database Administrator Exams. The question that I have is simple: “How relevant is this book because it was prepared for the older exams?” There isn’t a table of content published on the Amazon.com site. If there was a table of contents it could help me determine how close the book’s content is to the new exam.

As a preparation to figure out the value of the book as a study guide, I’ve reviewed the current Oracle MySQL Training Objectives (listed below). The new MySQL OCP Developer and Administrator exams have the following descriptions and objectives:

• MySQL 5.6 Developer 1Z0-882. Oracle provides the following outline for their MySQL for Developer (Ed 3) training course:

  Course Objectives

  • Describe the MySQL client/server architecture
  • Use MySQL client programs and common options
  • Program MySQL applications with Java and PHP connectors
  • Use a “NoSQL” approach to store and retrieve data
  • Design efficient tables
  • Create and delete database objects
  • Use expressions in SQL statements
  • Examine database metadata
  • Use SQL statements to modify table data
  • Maintain database integrity with transactions
  • Write multiple table queries
  • Create “virtual tables” containing specific data
  • Create user-defined variables, prepared statements, and stored routines
  • Create and manage triggers
  • Identify and deal with errors and exceptions in client programs
  • Write queries that are fast and effective, even under heavy loads
• MySQL 5.6 Database Administrator 1Z0-883. Oracle provides the following outline for their MySQL for Database Administrators (Ed 3.1) training course:

  Course Objectives

  • Describe the MySQL Architecture
  • Install and Upgrade MySQL
  • Use the INFORMATION_SCHEMA database to access metadata
  • Perform the MySQL start and shutdown operations
  • Configure MySQL server options at runtime
  • Use available tools including MySQL Workbench
  • Evaluate data types and character sets for performance issues
  • Understand data locking in MySQL
  • Understand the use of the InnoDB storage engine with MySQL
  • Maintain integrity of a MySQL installation
  • Use triggers for administration tasks
  • Use Enterprise Audit and Pluggable Authentication
  • Configure advanced replication topologies for high availability
  • Describe introductory performance tuning techniques
  • Perform backup and restore operations
  • Automate administrative tasks with scheduled events

  As always, I hope this helps those who read it; and, in this case I hope it helps you make an effective decision on preparation resources for the MySQL 5.6 OCP exams.