MacLochlainns Weblog

Michael McLaughlin's Technical Blog

Site Admin

Using CALIBRATE_IO

without comments

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.

Written by maclochlainn

August 31st, 2015 at 8:59 pm

Free Oracle Tuning Book

with one comment

Quick Start Guide to Oracle Query TuningWho 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).

Written by maclochlainn

August 31st, 2015 at 11:24 am

Use an object in a query?

without comments

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.

As always, I hope this helps those solving problems.

Written by maclochlainn

August 22nd, 2015 at 5:23 pm

ORDER BY CASE

with 6 comments

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.

Written by maclochlainn

July 8th, 2015 at 10:06 pm

Mac SQL Developer Install

without comments

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:

YosemiteInstallJava_01

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:

YosemiteInstallJava_02

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:

SQLDeveloperDownload

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:

SQLDeveloperInstall_01

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:

SQLDeveloperInstall_02

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

SQLDeveloperInstall_03

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

SQLDeveloperInstall_04

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

SQLDeveloperInstall_05

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

SQLDeveloperInstall_06

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

Written by maclochlainn

June 12th, 2015 at 3:08 am

Run X11 Apps on Mac

without comments

It’s possible folks didn’t notice but Mac OS X no longer includes XQuartz by default from Maverick forward. You need to download XQuartz and install it. I’d recommend after you install Xcode.

Launch XQuartz and then either use the bash shell it opens or open a Terminal bash shell session. Inside the shell, you might start Secure Shell (ssh) like this:

Mac-Pro-3:~ michaelmclaughlin$ ssh student@192.168.2.170
student@192.168.2.170's password: 
Last login: Thu Jun  4 14:33:37 2015
[student@localhost ~]$ xclock &
[1] 10422
[student@localhost ~]$ Error: Can't open display:

Granted that’s a trivial error and running the xclock X11 applications isn’t crucial, an error that makes it more important is the following from Oracle’s old Designer/2000 application:

FRM-91111: Internal Error: window system startup failure.
FRM-10039: Unable to start up the Form Builder.

This is the desired behavior. Secure shell (ssh) can’t run it unless you make the connection with the -Y flag. You should use the following syntax:

Mac-Pro-3:~ michaelmclaughlin$ ssh -Y student@192.168.2.170
student@192.168.2.170's password: 
Last login: Tue Jun  9 14:56:55 2015 from 192.168.2.1
/usr/bin/xauth:  file /home/student/.Xauthority does not exist
[student@localhost ~]$ xclock &
[1] 10760

You can safely ignore the .Xauthority does not exist warning message because it’ll create a .Xauthority file and store the magic cookie after the warning message. You should see the xclock program running in the upper left hand corner of your console, like:

X11MacXclock

It’s terrific that you don’t get a font warning like you typically would using UTF-8 on Linux. Nice that the Mac OS fonts are so well done that there isn’t a raised exception.

Using xclock or xeyes isn’t very useful as a rule, but this method also lets you run any of the Linux GUI applications. For example, the following gedit command lets you run the gedit utility from a Mac OS console. If you’ve installed the gedit plug-ins, you also can use the Terminal console on the remote system.

X11GeditTerminal

The process sequence for the command-line is shown below:

1030     1  /usr/sbin/sshd -D     - The root process launches the ssh daemon
3145  1030  sshd: student [priv]  - The sshd launches a ssh session to manage a student ssh session
3152  3145  sshd: student@pts/1   - The ssh session launched to manage the ssh session
3166  3152  -bash                 - The bash shell launched by connecting through the ssh session
3240  3166  gedit                 - The gedit command issued inside a ssh session
3166  3240  gnome-pty-helper      - Launching the gedit session across X11 
3169  3240  /bin/bash             - Launching the Terminal session inside the gedit session across X11
3269  3884  ps -ef                - Command run inside the gedit Terminal session

Hope that helps those who want to use X11 applications on the Mac OS.

Written by maclochlainn

June 9th, 2015 at 4:51 pm

Fedora X11 Install

without comments

While working through getting my Mac OS X to work with X11, I stumbled on some interesting errors and misdirection solutions. Like most things, the solution was straightforward. Then, it struck me that I hadn’t installed it on my Fedora image. This blog post show you the errors I got the way to get it to work, and how to install X11 on Fedora.

The first step requires discovering the package. If you remember xclock or xeyes are X-Windows programs, it’s quite easy with this command (though it may take a moment or two to run):

repoquery -q -f */xclock

It will return something like this:

xorg-x11-apps-0:7.7-7.fc20.x86_64

You can then install X11 as a sudoer user with the yum utility like this:

sudo yum -y install xorg-x11-apps

It should return this to your console:

Loaded plugins: langpacks, refresh-packagekit
Resolving Dependencies
--> Running transaction check
---> Package xorg-x11-apps.x86_64 0:7.7-7.fc20 will be installed
--> Processing Dependency: xorg-x11-xbitmaps for package: xorg-x11-apps-7.7-7.fc20.x86_64
--> Running transaction check
---> Package xorg-x11-xbitmaps.noarch 0:1.1.1-6.fc20 will be installed
--> Finished Dependency Resolution
 
Dependencies Resolved
 
================================================================================
 Package                  Arch          Version             Repository     Size
================================================================================
Installing:
 xorg-x11-apps            x86_64        7.7-7.fc20          fedora        305 k
Installing for dependencies:
 xorg-x11-xbitmaps        noarch        1.1.1-6.fc20        fedora         37 k
 
Transaction Summary
================================================================================
Install  1 Package (+1 Dependent package)
 
Total download size: 341 k
Installed size: 949 k
Downloading packages:
(1/2): xorg-x11-apps-7.7-7.fc20.x86_64.rpm                  | 305 kB  00:01     
(2/2): xorg-x11-xbitmaps-1.1.1-6.fc20.noarch.rpm            |  37 kB  00:00     
--------------------------------------------------------------------------------
Total                                              252 kB/s | 341 kB  00:01     
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction (shutdown inhibited)
  Installing : xorg-x11-xbitmaps-1.1.1-6.fc20.noarch                        1/2 
  Installing : xorg-x11-apps-7.7-7.fc20.x86_64                              2/2 
  Verifying  : xorg-x11-apps-7.7-7.fc20.x86_64                              1/2 
  Verifying  : xorg-x11-xbitmaps-1.1.1-6.fc20.noarch                        2/2 
 
Installed:
  xorg-x11-apps.x86_64 0:7.7-7.fc20                                             
 
Dependency Installed:
  xorg-x11-xbitmaps.noarch 0:1.1.1-6.fc20                                       
 
Complete!

After you install the xorg-x11-apps libraries, you can launch xclock. You should use the following syntax:

xclock &

It should display something like the following on your console:

X11xclock

The warning message is typically because you’re running something like en_US.UTF-8 mode. You can find suitable X11 character sets by using the following command:

sudo yum search xorg-x11-fonts

You can install all of them with the following command:

sudo yum -y install xorg-x11-fonts*

However, at the end of the day the warning doesn’t go way. You should just ignore it.

Hope this helps those who want to install X11 on Fedora.

Written by maclochlainn

June 9th, 2015 at 3:53 pm

Bash Arrays & Oracle

with 2 comments

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.

LittleERDModel

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.

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:

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.

Written by maclochlainn

May 21st, 2015 at 1:16 am

Bash Arrays & MySQL

without comments

Student questions are always interesting! They get me to think and to write. The question this time is: “How do I write a Bash Shell script to process multiple MySQL script files?” This post builds the following model (courtesy of MySQL Workbench) by using a bash shell script and MySQL script files, but there’s a disclaimer on this post. It shows both insecure and secure approaches and you should avoid the insecure ones.

LittleERDModel

It seems a quick refresher on how to use arrays in bash shell may be helpful. While it’s essential in a Linux environment, it’s seems not everyone masters the bash shell.

Especially, since I checked my Learning the Bash Shell (2nd Edition) and found a typo on how you handle arrays in the bash shell, and it’s a mistake that could hang newbies up (on page 161). Perhaps I should update my copy because I bought it in 1998. 😉 It was good then, and the new edition is probably better. The error is probably corrected in the current Learning the Bash Shell, but if not, the following examples show you how to use arrays in loops.

Naturally, these do presume some knowledge of working with bash shell, like the first line always is the same in any bash shell script. That you open an if-statement with an if and close it with a fi, and that you else-if is elif; and that a semicolon between a for-statement and the do statement is required when they’re on the same line because they’re two statements.

If you’re new to bash shell arrays, click on the link below to expand a brief tutorial. It takes you through three progressive examples of working with bash arrays.

Only one more trick needs to be qualified before our main MySQL examples. That trick is how you pass parameters to a bash shell script. For reference, this is the part that’s insecure because user command histories are available inside the Linux OS.

Here’s a hello_whom.sh script to demonstrates the concept of parameter passing:

1
2
3
4
5
6
7
8
9
10
#!/usr/bin/bash
 
# This says hello to the argument while managing no argument.
if [[ ${#} = 1 ]]; then
  echo 'The '${0}' program says: "Hello '${1}'!"'
elif [[ ${#} > 1 ]]; then
  echo 'The '${0}' program wants to know if you have more than one name?'
else
  echo 'The '${0}' program wants to know if you have a name?'
fi

If you need more on how parameters are passed and managed, you can check a prior blob post on Handling bash Parameters, or check the bash help pages. The following leverages bash arrays to run scripts and query the MySQL database from the command line.

You will need the three batch SQL files first, so here they are:

The following list_mysql.sh shell script expects to receive the username, password, database and fully qualified path in that specific order. The script names are entered manually because this should be a unit test script. Naturally, you can extend the script to manage those parameters but as mentioned I see this type of solution as a developer machine only script to simplify unit testing. Anything beyond that is risky!

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
#!/usr/bin/bash
 
# Assign user and password
username="${1}"
password="${2}"
database="${3}"
directory="${4}"
 
# List the parameter values passed.
echo "Username:  " ${username}
echo "Password:  " ${password}
echo "Database:  " ${database}
echo "Directory: " ${directory}
echo ""
 
# 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 ${cmd[*]}; do
  mysql -s -u${username} -p${password} -D${database} < ${directory}/${i} > /dev/null 2>/dev/null
done
 
# Connect and pipe the query result minus errors and warnings to the while loop.
mysql -u${username} -p${password} -D${database} <<<'show tables' 2>/dev/null |
 
# Read through the piped result until it's empty but format the title.
while IFS='\n' read list; do
  if [[ ${list} = "Tables_in_sampledb" ]]; then
    echo $list
    echo "----------------------------------------"
  else
    echo $list
  fi
done
echo ""
 
# Connect and pipe the query result minus errors and warnings to the while loop.
mysql -u${username} -p${password} -D${database} <<<'SELECT CONCAT(a.actor_name," in ",f.film_name) AS "Actors in Films" 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' 2>/dev/null |
 
# Read through the piped result until it's empty but format the title.
while IFS='\n' read actor_name; do
  if [[ ${actor_name} = "Actors in Films" ]]; then
    echo $actor_name
    echo "----------------------------------------"
  else
    echo $actor_name
  fi
done

The IFS (Internal Field Separator) works with whitespace by default. The IFS on lines 33 and 47 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 this script with the following input parameters from the local directory where you deploy it. The a parameters are: (1) username, (2) password, (3) database, and (4) a fully qualified path to the SQL setup files.

./list_mysql.sh student student sampledb "/home/student/Code/bash/mysql"

With valid input values, the list_mysql.sh bash script generates the following output, which confirms inputs and verifies actions taken by the scripts with queries:

Username:   student
Password:   student
Database:   sampledb
Directory:  /home/student/Code/bash/mysql
 
Tables_in_sampledb
----------------------------------------
actor
film
movie
 
Actors in Films
----------------------------------------
Chris Hemsworth in Thor
Chris Hemsworth in Thor: The Dark World
Chris Pine in Star Trek
Chris Pine in Star Trek into Darkness
Chris Pine in Guardians of the Galaxy

If you forgot to provide the required inputs to the list_mysql.sh bash script, it alternatively returns the following output:

Username:  
Password:  
Database:  
Directory: 
 
./list_mysql.sh: line 25: /actor.sql: No such file or directory
./list_mysql.sh: line 25: /film.sql: No such file or directory
./list_mysql.sh: line 25: /movie.sql: No such file or directory

The secure way removes the password at a minimum! The refactored program will require you to manually enter the password for all elements of the array (three in this sample), and twice for the two queries. Here’s the refactored code:

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
#!/usr/bin/bash
 
# Assign user and password
username="${1}"
database="${2}"
directory="${3}"
 
# List the parameter values passed.
echo "Username:  " ${username}
echo "Database:  " ${database}
echo "Directory: " ${directory}
echo ""
 
# 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 ${cmd[*]}; do
  mysql -s -u${username} -p -D${database} < ${directory}/${i} > /dev/null 2>/dev/null
done
 
# Connect and pipe the query result minus errors and warnings to the while loop.
mysql -u${username} -p -D${database} <<<'show tables' 2>/dev/null |
 
# Read through the piped result until it's empty.
while IFS='\n' read list; do
  if [[ ${list} = "Tables_in_sampledb" ]]; then
    echo $list
    echo "----------------------------------------"
  else
    echo $list
  fi
done
echo ""
 
# Connect and pipe the query result minus errors and warnings to the while loop.
mysql -u${username} -p -D${database} <<<'SELECT CONCAT(a.actor_name," in ",f.film_name) AS "Actors in Films" 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' 2>/dev/null |
 
# Read through the piped result until it's empty.
while IFS='\n' read actor_name; do
  if [[ ${actor_name} = "Actors in Films" ]]; then
    echo $actor_name
    echo "----------------------------------------"
  else
    echo $actor_name
  fi
done

Please let me know if you think there should be any more scaffolding for newbies in this post. As always, I hope this helps those looking for this type of solution.

Written by maclochlainn

May 17th, 2015 at 12:01 pm

Add Gedit Plugins

with one comment

Fedora comes with vim and gedit installed but the gedit installation is bare bones. You can update gedit to include supplemental Plug-ins with the following yum command as the root user:

yum install -y gedit-plugins

It generates the following log file:

Loaded plugins: langpacks, refresh-packagekit
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                                  |  14 kB  00:00     
updates                                                     | 4.9 kB  00:00     
(1/2): pgdg93/20/x86_64/primary_db                          |  86 kB  00:00     
(2/2): updates/20/x86_64/primary_db                         |  11 MB  00:03     
(1/2): updates/20/x86_64/pkgtags                            | 1.5 MB  00:00     
(2/2): updates/20/x86_64/updateinfo                         | 2.0 MB  00:01     
Resolving Dependencies
--> Running transaction check
---> Package gedit-plugins.x86_64 0:3.10.1-1.fc20 will be installed
--> Processing Dependency: libgit2-glib for package: gedit-plugins-3.10.1-1.fc20.x86_64
--> Running transaction check
---> Package libgit2-glib.x86_64 0:0.0.6-2.fc20 will be installed
--> Processing Dependency: libgit2.so.0()(64bit) for package: libgit2-glib-0.0.6-2.fc20.x86_64
--> Running transaction check
---> Package libgit2.x86_64 0:0.19.0-2.fc20 will be installed
--> Processing Dependency: libxdiff.so.1()(64bit) for package: libgit2-0.19.0-2.fc20.x86_64
--> Processing Dependency: libhttp_parser.so.2()(64bit) for package: libgit2-0.19.0-2.fc20.x86_64
--> Running transaction check
---> Package http-parser.x86_64 0:2.0-5.20121128gitcd01361.fc20 will be installed
---> Package libxdiff.x86_64 0:1.0-3.fc20 will be installed
--> Finished Dependency Resolution
 
Dependencies Resolved
 
================================================================================
 Package          Arch      Version                            Repository  Size
================================================================================
Installing:
 gedit-plugins    x86_64    3.10.1-1.fc20                      updates    830 k
Installing for dependencies:
 http-parser      x86_64    2.0-5.20121128gitcd01361.fc20      fedora      23 k
 libgit2          x86_64    0.19.0-2.fc20                      fedora     281 k
 libgit2-glib     x86_64    0.0.6-2.fc20                       fedora      82 k
 libxdiff         x86_64    1.0-3.fc20                         fedora      33 k
 
Transaction Summary
================================================================================
Install  1 Package (+4 Dependent packages)
 
Total download size: 1.2 M
Installed size: 5.2 M
Downloading packages:
(1/5): http-parser-2.0-5.20121128gitcd01361.fc20.x86_64.rpm |  23 kB  00:00     
(2/5): libgit2-0.19.0-2.fc20.x86_64.rpm                     | 281 kB  00:00     
(3/5): libgit2-glib-0.0.6-2.fc20.x86_64.rpm                 |  82 kB  00:00     
(4/5): libxdiff-1.0-3.fc20.x86_64.rpm                       |  33 kB  00:00     
(5/5): gedit-plugins-3.10.1-1.fc20.x86_64.rpm               | 830 kB  00:01     
--------------------------------------------------------------------------------
Total                                              899 kB/s | 1.2 MB  00:01     
Running transaction check
Running transaction test
Transaction test succeeded
Running transaction (shutdown inhibited)
  Installing : libxdiff-1.0-3.fc20.x86_64                                   1/5 
  Installing : http-parser-2.0-5.20121128gitcd01361.fc20.x86_64             2/5 
  Installing : libgit2-0.19.0-2.fc20.x86_64                                 3/5 
  Installing : libgit2-glib-0.0.6-2.fc20.x86_64                             4/5 
  Installing : gedit-plugins-3.10.1-1.fc20.x86_64                           5/5 
  Verifying  : libgit2-0.19.0-2.fc20.x86_64                                 1/5 
  Verifying  : libgit2-glib-0.0.6-2.fc20.x86_64                             2/5 
  Verifying  : gedit-plugins-3.10.1-1.fc20.x86_64                           3/5 
  Verifying  : http-parser-2.0-5.20121128gitcd01361.fc20.x86_64             4/5 
  Verifying  : libxdiff-1.0-3.fc20.x86_64                                   5/5 
 
Installed:
  gedit-plugins.x86_64 0:3.10.1-1.fc20                                          
 
Dependency Installed:
  http-parser.x86_64 0:2.0-5.20121128gitcd01361.fc20                            
  libgit2.x86_64 0:0.19.0-2.fc20                                                
  libgit2-glib.x86_64 0:0.0.6-2.fc20                                            
  libxdiff.x86_64 0:1.0-3.fc20                                                  
 
Complete!

When you launch the gedit utility, you click on the

Gedit Plug-in Installation

GeditPref_01

  1. After you install the Gedit Plug-ins, you can configure the plug-ins by launching Gedit and then click on the gedit menu option. Then, click on the Preferences menu option to enable the new plugins, like the Embedded Terminal plug-in.

GeditPref_02

  1. You have four tab options when working with the Preferences menu. The first tab is the View tab, as shown to the left.

GeditPref_03

  1. The second tab is the Editor tab, as shown to the left.

GeditPref_04

  1. The third tab is the Font & Colors tab, as shown to the left.

GeditPref_05

  1. The fourth tab is the Plugins tab, as shown to the left. Scroll down the list and check the Embedded Terminal and Python Console plug-ins’ checkbox. The Embedded Terminal lets you edit a file and have command line access to a Terminal session from the gedit menu; and the Python Console session from the gedit menu.

GeditPref_06

  1. Click on the View menu, and then choose the Bottom Panel menu option.

GeditPref_07

  1. After enabling the Bottom Panel in the Gedit menu, you can edit a file and click on the Terminal by simply clicking on the subpanel. You can see the split image on the left. There’s also a set of bottom tabs that lets you switch from a Linux Terminal session to the Python console.

As always, I hope this helps those working with gedit on the Fedora operating system.

Written by maclochlainn

May 15th, 2015 at 1:23 am