Archive for the ‘Oracle DBA’ tag
REGEXP_LIKE Behavior
Often, the biggest problem with regular expressions is that those who use them sometimes don’t use them correctly. A great example occurs in the Oracle Database with the REGEXP_LIKE
function. For example, some developer use the following to validate whether a string is a number but it only validates whether the first character is a number.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 | DECLARE lv_input VARCHAR2(100); BEGIN /* Assign input value. */ lv_input := '&input'; /* Check for numeric string. */ IF REGEXP_LIKE(lv_input,'[[:digit:]]') THEN dbms_output.put_line('It''s a number.'); ELSE dbms_output.put_line('It''s a string.'); END IF; END; / |
When they test numbers it appears to works, it even appears to work when the test string start with number, but it fails with any string that starts with a character. That’s because the REGEXP_LIKE
function on line 8 only checks the first character, but the following checks all the characters in the string.
8 | IF REGEXP_LIKE(lv_inputs(i),'[[:digit:]]{'||LENGTH(lv_inputs(i))||'}') THEN |
You can also fix it with the following non-Posix solution:
8 | IF REGEXP_LIKE(lv_input,'[[0-9]]') THEN |
You can add a collection to the program and use it to test single-digit, double-digit, and string with a leading integer. Save the program as test.sql
and you can test three conditions with one call.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 | DECLARE /* Declare the local collection type. */ TYPE inputs IS TABLE OF VARCHAR2(100); /* Declare a local variable of the collection type. */ lv_inputs INPUTS; BEGIN /* Assign the inputs to the collection variable. */ lv_inputs := inputs('&1','&2','&3'); /* Read through the collection and print whether it's an number or string. */ FOR i IN 1..lv_inputs.COUNT LOOP IF REGEXP_LIKE(lv_inputs(i),'[[:digit:]]{'||LENGTH(lv_inputs(i))||',}') THEN dbms_output.put_line('It''s a number.'); ELSE dbms_output.put_line('It''s a string.'); END IF; END LOOP; END; / |
You can run the test.sql
program like this:
SQL> @test.sql 1 12 1a |
It prints:
It's a number. It's a number. It's a string. |
As always, I hope this helps those looking for a solution.
SQL*Plus Tricks
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 CALIBRATE_IO
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.
Free Oracle Tuning Book
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:
- Query Tuning: Developer and Beginning DBA
- Query Tuning: Basics for DBAs and Developers
- Advanced Performance Tuning
- 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).
Use an object in a query?
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.
Setup Object Types ↓
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.
Bash Arrays & Oracle
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.
Setup SQL Files ↓
The actor.sql
file:
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; |
The film.sql
file:
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; |
The movie.sql
file:
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; |
The tables.sql
file, lets you verify the creation of the actor
, film
, and movie
tables:
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; |
The results.sql
file, lets you see join results from actor
, film
, and movie
tables:
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:
Create sample
User SQL File ↓
You can use the dbms_metadata.get_ddl
function to discover the existence of the tablespaces. The following SQL syntax returns the SQL DDL statement that created a users
or temp
tablespace:
1 2 | SET LONG 200000 SELECT dbms_metadata.get_ddl('TABLESPACE','USERS') FROM dual; |
You create the sample
database with the following SQL statements:
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.
Leaf node queries
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.
SQL Developer – Fedora
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:
Running this product is supported with a minimum Java version of 1.7.0_51 and a maximum version less than 1.8.
Update the SetJavaHome in “/root/.sqldeveloper/4.0.0/product.conf” to point to another Java.
This produce will not be supported, and may not run correctly if you proceed. Continue anyway?
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.
APEX Create Table
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.
- 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 replacelocalhost
with the IP address orhostname
for the IP address.- Workspace:
STUDENT
- Username:
ADMIN
- Password:
STUDENT
- Workspace:
- 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 athttp://localhost:8080/apex
by default on the server. If you’ve got a static IP address for your instance, you can replacelocalhost
with the IP address orhostname
for the IP address. Click on the Object Browser icon to proceed.
- Clicking the SQL Workshop icon takes you to the second level menu. You click the Object Browser icon to create a database object.
- After clicking the Object Browser icon, you see the screen at the left. Click the Create button to create a table.
- 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.
- 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.
- 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 applyingNOT 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.
- 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 theIMAGE_ID
column, and declares anIMAGE_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.
- 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.
- 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.
- 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.
- 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.
- 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.
- 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") /
- 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.
APEX SQL Query
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.
- 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 replacelocalhost
with the IP address orhostname
for the IP address.- Workspace:
STUDENT
- Username:
ADMIN
- Password:
STUDENT
- Workspace:
- 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 athttp://localhost:8080/apex
by default on the server. If you’ve got a static IP address for your instance, you can replacelocalhost
with the IP address orhostname
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.
- 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.
- The first text panel lets you enter free-form queries. The Autocommit checkbox is enabled, which means the result of
INSERT
andUPDATE
statements are immediate and don’t require aCOMMIT
statement. The second text panel displays results from a query or acknowledgment of statement completion.
- 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.