MacLochlainns Weblog

Michael McLaughlin's Technical Blog

Site Admin

Archive for the ‘DevOps’ Category

Oracle 23c Free SQL*Plus

without comments

It’s always frustrated me when using the sqlplus command-line interface (CLI) that you can’t just “up arrow” to through the history. At least, that’s the default case unless you wrap the sqlplus executable.

I like to do my development work as close to the database as possible. The delay from SQL Developer to the database or VSCode to the database is just too long. Therefore, I like the native sqlplus to be as efficient as possible. This post shows you how to install the rlwarp utility to wrap sqlplus and create a sandboxed student user for a local development account inside the Oracle 23c Free container. You should note that the Docker or Podman Container is using Oracle Unbreakable Linux 8 as it’s native OS.

You can connect to your Docker version of Oracle Database 23c Free with the following command:

docker exec -it -u root oracle23c bash

You can’t just use dnf to install rlwrap and get it to magically install all the dependencies. That would be too easy, eh?

Attempting to do so will lock your base OS and eventually force you to kill with prejudice the hung dnf process (at least it forced me to do so). You need to determine the rlwrap dependencies and then install them first. In that process, I noticed that the which utility program wasn’t installed in the container.

Naturally, I installed the which utility first with this command:

dnf install -y which

The rlwrap dependencies are: glibc, ncurses, perl, readline, python, and git. Only the perl, python, and git are missing from the list of formal dependencies but there’s another dependency the epel-release package.

If you want to verify whether a package is installed, you can use the rpm command like this:

rpm -qa | grep package_name

I installed the perl programming environment (a big install) with this command:

dnf install -y perl

I installed the python3 with this command:

dnf install -y python3

I installed the git module with this command:

dnf install -y git

I installed the epel-release container with this command:

dnf install -y epel-release

After installing all of these, you’re now ready to install the core rlwrap utility program. Like the other installations, you use:

dnf install -y rlwrap

At this point, you need to create a sandboxed user account for the Docker instance because as a developer using the root user for simple tasks is a bad idea. While you could do this with a Docker command, the Oracle 23c Free edition raised a lock on the /etc/group file when I tried it. Naturally, that’s not a problem because you can connect as the root user with this syntax:

docker exec -it -u root oracle23c bash

As the root user, create a student account as a developer account in the Oracle 23c Free container:

useradd -u 501 -g dba -G users -d /home/student -s /bin/bash/ -c "Student" -n student

You’ll be unable to leverage the tnsnames.ora file unless you alter the prior command to replace dba with oinstall or add the following command:

usermod -a -G oinstall student

Exit the Oracle 23c Free container as the root user and reconnect as the student user with this syntax:

docker exec -it --user student oracle23c bash

While you’re connected as the root user, you should create an upload directory as a subdirectory of the $ORACLE_BASE directory. The $ORACLE_BASE directory in the Oracle Database 23c Free Docker image is the /opt/oracle directory.

You should use the following syntax to create the upload directory and change its permission to that of the Oracle Database 23c Free installation (for a future blog post on developing external table deployment on the Docker image):

mkdir /opt/oracle
chown -R oracle:install /opt/oracle/upload

You also can add the following student function to the Ubuntu student user’s .bashrc file. It means all you need to type to connect to the Oracle Database 23c Free Docker instance is “student“. I like shortcuts like this one, which let you leverage one-line Python commands.

student () 
{
    # Discover the fully qualified program name. 
    path=`which docker 2>/dev/null`
    file=''
 
    # Parse the program name from the path.
    if [ -n ${path} ]; then
        file=${path##/*/}
    fi
 
    # Wrap when there is a file and it is rewrap.
    if [ -n ${file} ] && [[ ${file} = "docker" ]]; then
        python -c "import subprocess; subprocess.run(['docker exec -it --user student oracle23c bash'], shell=True)" 
    else
        echo "Docker is unavailable: Install the docker package."
    fi
}

Open a Ubuntu Terminal shell and type a student function name to connect to the Docker Oracle Database 23c Free instance where you can now test things like external tables with the SQL*Plus command line without installing it on the Ubuntu local operating system.

student@student-virtual-machine:~$ student
[student@d28375f0c43f ~]$ sqlplus c##student/student@free
 
SQL*Plus: Release 23.0.0.0.0 - Production on Wed Jan 3 02:14:22 2024
Version 23.3.0.23.09
 
Copyright (c) 1982, 2023, Oracle.  All rights reserved.
 
Last Successful login time: Wed Jan 03 2024 01:56:44 +00:00
 
Connected to:
Oracle Database 23c Free Release 23.0.0.0.0 - Develop, Learn, and Run for Free
Version 23.3.0.23.09
 
SQL>

Then, I added this sqlplus function to the /home/student/.bashrc file, which is owned by the student user. However, I also added the instruction to change to the student user’s home directory because the Oracle 23c Free container will connect you to the /home/oracle directory by default. I also added the default long list (ll) alias to the .bashrc file.

sqlplus () 
{
    # Discover the fully qualified program name. 
    path=`which rlwrap 2>/dev/null`
    file=''
 
    # Parse the program name from the path.
    if [ -n ${path} ]; then
        file=${path##/*/}
    fi;
 
    # Wrap when there is a file and it is rewrap.
    if [ -n ${file} ] && [[ ${file} = "rlwrap" ]]; then
        rlwrap sqlplus "${@}"
    else
        echo "Command-line history unavailable: Install the rlwrap package."
        $ORACLE_HOME/bin/sqlplus "${@}"
    fi
}
 
# Change to the user's home directory.
cd ${HOME}
 
# Create a long list alias:
alias ll='ls -l --color=auto'

After you’ve configured your student user, you can configure the oracle user account to work like a regular server. Exit the Docker Oracle Database 23c Free as the student user, then connect as the root user with this command:

docker exec -it -u root oracle23c bash

As the root user you can become the oracle user with the following command:

su - oracle

Now, add the following .bashrc shell in the /home/oracle directory:

# The oracle user's .bashrc
 
# Source global definitions
if [ -f /etc/bashrc ]; then
	. /etc/bashrc
fi
 
# User specific environment
if ! [[ "$PATH" =~ "$HOME/.local/bin:$HOME/bin:" ]]
then
    PATH="$HOME/.local/bin:$HOME/bin:$PATH"
fi
export PATH
 
# Uncomment the following line if you don't like systemctl's auto-paging feature:
# export SYSTEMD_PAGER=
 
# User specific aliases and functions
export ORACLE_SID=FREE
export ORACLE_BASE=/opt/oracle
export ORACLE_HOME=/opt/oracle/product/23c/dbhomeFree
export PATH=$PATH:/$ORACLE_HOME/bin
 
# Change to the user's home directory.
cd ${HOME}
 
# Create a long list alias:
alias ll='ls -l --color=auto'
 
sqlplus () 
{
    # Discover the fully qualified program name. 
    path=`which rlwrap 2>/dev/null`
    file=''
 
    # Parse the program name from the path.
    if [ -n ${path} ]; then
        file=${path##/*/}
    fi;
 
    # Wrap when there is a file and it is rewrap.
    if [ -n ${file} ] && [[ ${file} = "rlwrap" ]]; then
        rlwrap sqlplus "${@}"
    else
        echo "Command-line history unavailable: Install the rlwrap package."
        $ORACLE_HOME/bin/sqlplus "${@}"
    fi
}

You need to manually source the .bashrc for the oracle user because it’s not an externally available user. Use this syntax to connect as the internal user:

sqlplus / as sysdba

It’ll display:

SQL*Plus: RELEASE 23.0.0.0.0 - Production ON Wed Jan 3 07:08:11 2024
Version 23.3.0.23.09
 
Copyright (c) 1982, 2023, Oracle.  ALL rights reserved.
 
 
Connected TO:
Oracle DATABASE 23c Free RELEASE 23.0.0.0.0 - Develop, Learn, AND Run FOR Free
Version 23.3.0.23.09
 
SQL>

After all this, I can now click the “up arrow” to edit any of the sqlplus command history. If you like to work inside sqlplus natively, this should help you.

Written by maclochlainn

December 20th, 2023 at 11:11 pm

Disk Space Allocation

without comments

It’s necessary to check for adequate disk space on your Virtual Machine (VM) before installing Oracle 23c Free in a Docker container or as a podman service. Either way, it requires about 13 GB of disk space. On Ubuntu, the typical install of a VM allocates 20 GB and a 500 MB swap. You need to create a 2 GB swap when you install Ubuntu or plan to change the swap, as qualified in this excellent DigitalOcean article. Assuming you installed it with the correct swap or extended your swap area, you can confirm it with the following command:

sudo swapon --show

It should return something like this:

NAME      TYPE SIZE USED PRIO
/swapfile file 2.1G 1.2G   -2

Next, check your disk space allocation and availability with this command:

df -h

This is what was in my instance with MySQL and PostgreSQL databases already installed and configured with sandboxed schemas:

Filesystem      Size  Used Avail Use% Mounted on
tmpfs           388M  2.1M  386M   1% /run
/dev/sda3        20G   14G  4.6G  75% /
tmpfs           1.9G   28K  1.9G   1% /dev/shm
tmpfs           5.0M  4.0K  5.0M   1% /run/lock
/dev/sda2       512M  6.1M  506M   2% /boot/efi
tmpfs           388M  108K  388M   1% /run/user/1000

Using VMware Fusion on my Mac (Intel-based i9), I changed the allocated space from 20 GB to 40 GB by navigating to Virtual Machine, Settings…, Hard Disk. I entered 40.00 as the disk size and clicked the Pre-allocate disk space checkbox before clicking the Apply button, as shown in below. This added space is necessary because Oracle Database 23c Free as a Docker instance requires almost 10 GB of local space.

After clicking the Apply button, I checked Ubuntu with the “df -h” command and found there was no change. That’s unlike doing the same thing on AlmaLinux or a RedHat distribution, which was surprising.

The next set of steps required that I manually add the space to the Ubuntu instance:

  1. Start the Ubuntu VM and check the instance’s disk information with fdisk:

    sudo fdisk -l

    The log file for this is:

    After running fdisk, I rechecked disk allocation with df -h and saw no change:

    Filesystem      Size  Used Avail Use% Mounted on
    tmpfs           388M  2.1M  386M   1% /run
    /dev/sda3        20G   14G  4.6G  75% /
    tmpfs           1.9G   28K  1.9G   1% /dev/shm
    tmpfs           5.0M  4.0K  5.0M   1% /run/lock
    /dev/sda2       512M  6.1M  506M   2% /boot/efi
    tmpfs           388M  108K  388M   1% /run/user/1000
  2. So, I installed Ubuntu’s user space utility gparted:

    sudo apt install gparted

    The log file for this is:

  3. After installing the gparted utility (manual can be found here), you can launch it with the following syntax:

    sudo apt install gparted

    You’ll see the following in the console, which you can ignore.

    GParted 1.3.1
    configuration --enable-libparted-dmraid --enable-online-resize
    libparted 3.4

    It launches a GUI interface that should look something like the following:

    Right-click on the /dev/sda3 Partition and the GParted application will present the following context popup menu. Click the Resize/Move menu option.

    The attempt to resize the disk at this point GParted will raise a read-only exception like the following:

    You might open a new shell and fix the disk at the command-line but you’ll need to relaunch gparted regardless. So, you should close gparted and run the following commands:

    sudo mount -o remount -rw /
    sudo mount -o remount -rw /var/snap/firefox/common/host-hunspell

    When you relaunch GParted, you see that the graphic depiction has changed when you right-click on the /dev/sda3 Partition as follows:

    Click on the highlighted box with the arrow and drag it all the way to the right. It will then show you something like the following.

    Click the Resize button to make the change and add the space to the Ubuntu file system and see something like the following in Gparted:

    Choose Edit in the menu bar and then Apply All Operations to effect the change in the disk allocation. The last dialog will require you to verify you want to make the changes. Click the Apply button to make the changes.

    Click the close for the GParted application and then you can rerun the following command:

    df -h

    You will see that you now have 19.5 GB of additional space:

    Filesystem      Size  Used Avail Use% Mounted on
    tmpfs           388M  2.2M  386M   1% /run
    /dev/sda3        39G 19.5G   23G  39% /
    tmpfs           1.9G   28K  1.9G   1% /dev/shm
    tmpfs           5.0M  4.0K  5.0M   1% /run/lock
    /dev/sda2       512M  6.1M  506M   2% /boot/efi
    tmpfs           388M  116K  388M   1% /run/user/1000
  4. Finally, you can now successfully download the latest Docker version of Oracle Database 23c Free with the following command:

    docker run --name oracle23c -p 1521:1521 -p 5500:5500 -e ORACLE_PWD=cangetin container-registry.oracle.com/database/free:latest

    Since you haven’t downloaded the container, you’ll get a warning that it is unable to find the image before it discovers it and downloads it. This will take several minutes. At the conclusion, it will start the Oracle Database Net Listener and begin updating files. the updates may take quite a while to complete.

    The basic download console output looks like the following and if you check your disk space you’ve downloaded about 14 GB in the completed container.

    Unable to find image 'container-registry.oracle.com/database/free:latest' locally
    latest: Pulling from database/free
    089fdfcd47b7: Pull complete 
    43c899d88edc: Pull complete 
    47aa6f1886a1: Pull complete 
    f8d07bb55995: Pull complete 
    c31c8c658c1e: Pull complete 
    b7d28faa08b4: Pull complete 
    1d0d5c628f6f: Pull complete 
    db82a695dad3: Pull complete 
    25a185515793: Pull complete 
    Digest: sha256:5ac0efa9896962f6e0e91c54e23c03ae8f140cf6ed43ca09ef4354268a942882
    Status: Downloaded newer image for container-registry.oracle.com/database/free:latest

    My detailed log file for the complete recovery operation is:

  5. You can connect to the Oracle Database 23c Free container with the following syntax:

    docker exec -it -u root oracle23c bash

    At the command-line, you connect to the Oracle Database 23c Free container with the following syntax:

    sqlplus system/cangetin@free

    You have arrived at the Oracle SQL prompt:

    SQL*Plus: Release 23.0.0.0.0 - Production on Fri Dec 1 00:13:55 2023
    Version 23.3.0.23.09
     
    Copyright (c) 1982, 2023, Oracle.  All rights reserved.
     
    Last Successful login time: Thu Nov 30 2023 23:27:54 +00:00
     
    Connected to:
    Oracle Database 23c Free Release 23.0.0.0.0 - Develop, Learn, and Run for Free
    Version 23.3.0.23.09
     
    SQL>

As always, I hope this helps those trying to work with the newest Oracle stack.

Written by maclochlainn

December 1st, 2023 at 3:08 pm

Listener for APEX

without comments

Unless dbca lets us build the listener.ora file, we often leave off some component. For example, running listener control program the following status indicates an incorrectly configured listener.ora file.

lsnrctl status

It returns the following, which displays an endpoint for the XDB Server (I’m using Oracle Database 11g XE because it’s pre-containerized and has a small testing footprint):

LSNRCTL for Linux: Version 11.2.0.2.0 - Production on 24-MAR-2023 00:59:06
 
Copyright (c) 1991, 2011, Oracle.  All rights reserved.
 
Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC_FOR_XE)))
STATUS of the LISTENER
------------------------
Alias                     LISTENER
Version                   TNSLSNR for Linux: Version 11.2.0.2.0 - Production
Start Date                21-MAR-2023 21:17:37
Uptime                    2 days 3 hr. 41 min. 29 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Default Service           XE
Listener Parameter File   /u01/app/oracle/product/11.2.0/xe/network/admin/listener.ora
Listener Log File         /u01/app/oracle/diag/tnslsnr/localhost/listener/alert/log.xml
Listening Endpoints Summary...
  (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC_FOR_XE)))
  (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=localhost)(PORT=1521)))
  (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=localhost)(PORT=8080))(Presentation=HTTP)(Session=RAW))
Services Summary...
Service "PLSExtProc" has 1 instance(s).
  Instance "PLSExtProc", status UNKNOWN, has 1 handler(s) for this service...
Service "XE" has 1 instance(s).
  Instance "XE", status READY, has 1 handler(s) for this service...
Service "XEXDB" has 1 instance(s).
  Instance "XE", status READY, has 1 handler(s) for this service...
The command completed successfully

The listener is missing the second SID_LIST_LISTENER value of CLRExtProc value. A complete listener.ora file should be as follows for the Oracle Database XE:

# listener.ora Network Configuration FILE:
 
SID_LIST_LISTENER =
  (SID_LIST =
    (SID_DESC =
      (SID_NAME = PLSExtProc)
      (ORACLE_HOME = /u01/app/oracle/product/11.2.0/xe)
      (PROGRAM = extproc)
    )
    (SID_DESC =
      (SID_NAME = CLRExtProc)
      (ORACLE_HOME = /u01/app/oracle/product/11.2.0/xe)
      (PROGRAM = extproc)
    )
  )
 
LISTENER =
  (DESCRIPTION_LIST =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC_FOR_XE))
      (ADDRESS = (PROTOCOL = TCP)(HOST = localhost.localdomain)(PORT = 1521))
    )
  )
 
DEFAULT_SERVICE_LISTENER = (XE)

With this listener.ora file, the Oracle listener control utility will return the following correct status, which hides the XDB Server’s endpoint:

LSNRCTL for Linux: Version 11.2.0.2.0 - Production on 24-MAR-2023 02:38:57
 
Copyright (c) 1991, 2011, Oracle.  All rights reserved.
 
Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC_FOR_XE)))
STATUS of the LISTENER
------------------------
Alias                     LISTENER
Version                   TNSLSNR for Linux: Version 11.2.0.2.0 - Production
Start Date                24-MAR-2023 02:38:15
Uptime                    0 days 0 hr. 0 min. 42 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Default Service           XE
Listener Parameter File   /u01/app/oracle/product/11.2.0/xe/network/admin/listener.ora
Listener Log File         /u01/app/oracle/product/11.2.0/xe/log/diag/tnslsnr/localhost/listener/alert/log.xml
Listening Endpoints Summary...
  (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC_FOR_XE)))
  (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=localhost)(PORT=1521)))
Services Summary...
Service "CLRExtProc" has 1 instance(s).
  Instance "CLRExtProc", status UNKNOWN, has 1 handler(s) for this service...
Service "PLSExtProc" has 1 instance(s).
  Instance "PLSExtProc", status UNKNOWN, has 1 handler(s) for this service...
The command completed successfully

It seems a number of examples on the web left the SID_LIST_LISTENER value of CLRExtProc value out of the listener.ora file. As always, I hope this helps those looking for a complete solution rather than generic instructions without a concrete example.

Written by maclochlainn

March 24th, 2023 at 1:00 am

AWS EC2 TNS Listener

without comments

Having configured an AlmaLinux 8.6 with Oracle Database 11g XE, MySQL 8.0.30, and PostgreSQL 15, we migrated it to AWS EC2 and provisioned it. We used the older and de-supported Oracle Database 11g XE because it didn’t require any kernel modifications and had a much smaller footprint.

I had to address why attempting to connect with the sqlplus utility raised the following error after provisioning a copy with a new static IP address:

ERROR:
ORA-12514: TNS:listener does NOT currently know OF service requested IN CONNECT descriptor

A connection from SQL Developer raises a more addressable error, like:

ORA-17069

I immediately tried to check the connection with the tnsping utility and found that tnsping worked fine. However, when I tried to connect with the sqlplus utility it raised an ORA-12514 connection error.

There were no diagnostic steps beyond checking the tnsping utility. So, I had to experiment with what might block communication.

I changed the host name from ip-172-58-65-82.us-west-2.compute.internal to a localhost string in both the listener.ora and tnsnames.ora. The listener.ora file:

# listener.ora Network Configuration FILE:
 
SID_LIST_LISTENER =
  (SID_LIST =
    (SID_DESC =
      (SID_NAME = PLSExtProc)
      (ORACLE_HOME = /u01/app/oracle/product/11.2.0/xe)
      (PROGRAM = extproc)
    )
  )
 
LISTENER =
  (DESCRIPTION_LIST =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC_FOR_XE))
      (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))
    )
  )
 
DEFAULT_SERVICE_LISTENER = (XE)

The tnsnames.ora file:

# tnsnames.ora Network Configuration FILE:
 
XE =
  (DESCRIPTION =
    (ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))
    (CONNECT_DATA =
      (SERVER = DEDICATED)
      (SERVICE_NAME = XE)
    )
  )
 
EXTPROC_CONNECTION_DATA =
  (DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC_FOR_XE))
    )
    (CONNECT_DATA =
      (SID = PLSExtProc)
      (PRESENTATION = RO)
    )
  )

I suspected that it might be related to the localhost value. So, I checked the /etc/hostname and /etc/hosts files.

Then, I modified /etc/hostname file by removing the AWS EC2 damain address. I did it on a memory that Oracle’s TNS raises errors for dots or periods in some addresses.

The /etc/hostname file:

ip-172-58-65-82

The /etc/hosts file:

127.0.0.1   localhost localhost.localdomain localhost4 localhost4.localdomain4 ip-172-58-65-82
::1         localhost localhost.localdomain localhost6 localhost6.localdomain6 ip-172-58-65-82

Now, we can connect to the Oracle Database 11g XE instance with the sqlplus utility. I believe this type of solution will work for other AWS EC2 provisioned Oracle databases.

Written by maclochlainn

March 22nd, 2023 at 10:09 pm

AlmaLinux MySQL+Perl

without comments

A quick primer on Perl programs connecting to the MySQL database. It’s another set of coding examples for the AlmaLinux instance that I’m building for students. This one demonstrates basic Perl programs, connecting to MySQL, returning data sets by reference and position, dynamic queries, and input parameters to dynamic queries.

  1. Naturally, a hello.pl is a great place to start:

    #!/usr/bin/perl
     
    # Hello World program.
    print "Hello World!\n";

    After setting the permissions to -rwxr-xr-x. with this command:

    chmod 755 hello.pl

    You call it like this from the Command-Line Interface (CLI):

    ./hello.pl

    It prints:

    Hello World!
  2. Next, a connect.pl program lets us test the Perl::DBI connection to the MySQL database.

    #!/usr/bin/perl
     
    # Import libraries.
    use strict;
    use warnings;
    use v5.10;     # for say() function
    use DBI;
     
    # Print with say() function message.
    say "Perl MySQL Connect Attempt.";
     
    # MySQL data source name should have a valid database as the
    # third argument; this uses the sakila demo database.
    my $dsn = "DBI:mysql:sakila";
     
    # Local variables to build the connection string.
    my $username = "student";
    my $password = "student";
     
    # Set arguments for MySQL database error management.
    my %attr = ( PrintError=>0,  # turn off error reporting via warn()
                 RaiseError=>1);   # turn on error reporting via die()           
     
    # Create connction with a data source name, user name and password.
    my $dbh  = DBI->connect($dsn,$username,$password, \%attr);
     
    # Print with say() function valid connection message.
    say "Connected to the MySQL database.";

    After setting the permissions to -rwxr-xr-x. you call it with this CLI command:

    ./connect.pl

    It prints:

    Perl MySQL Connect Attempt.
    Connected to the MySQL database.
  3. After connecting to the database lets query a couple columns by reference notation in a static.pl program. This one just returns the result of the MySQL version() and database() functions.

    #!/usr/bin/perl
     
    # Import libraries.
    use strict;
    use warnings;
    use v5.10;     # for say() function
    use DBI;
     
    # Print with say() function message.
    say "Perl MySQL Connect Attempt.";
     
    # MySQL data source name must have a valid database as the
    # third argument; this uses the sakila demo database.
    my $dsn = "DBI:mysql:sakila";
     
    # Local variables to build the connection string.
    my $username = "student";
    my $password = "student";
     
    # Set arguments for MySQL database error management.
    my %attr = ( PrintError=>0,  # turn off error reporting via warn()
                 RaiseError=>1); # turn on error reporting via die()           
     
    # Create connction with a data source name, user name and password.
    my $dbh  = DBI->connect($dsn,$username,$password, \%attr);
     
    # Creaet a static SQL statement or query.
    my $sth = $dbh->prepare("SELECT version() AS version \
    	                 ,      database() AS db_name");
     
    # Execute the static statement.
    $sth->execute() or die "Execution failed: $dbh->errstr()";
     
    # Read data and print by reference.
    print "----------------------------------------\n";
    while (my $ref = $sth->fetchrow_hashref()) {
      print "MySQL Version:  $ref->{'version'}\nMySQL Database: $ref->{'db_name'}\n";
    }
    print "----------------------------------------\n";
     
    # Close the statement.
    $sth->finish;
     
    # Disconnect from database connection.
    $dbh->disconnect();
     
    # Print with say() function valid connection message.
    say "Connected to the MySQL database.";

    After setting the permissions to -rwxr-xr-x. you call it with this CLI command:

    ./static.pl

    It prints:

    Perl MySQL Connect Attempt.
    ----------------------------------------
    MySQL Version:  8.0.30
    MySQL Database: sakila
    ----------------------------------------
    Connected to the MySQL database.
  4. After connecting to the database and securing variables by reference notation, lets return the value as an array of rows in a columns.pl program. This one just returns data from the film table of the sakila database. It is a static query because all the values are contained inside the SQL statement.

    #!/usr/bin/perl
     
    # Import libraries.
    use strict;
    use warnings;
    use v5.10;     # for say() function
    use DBI;
     
    # Print with say() function message.
    say "Perl MySQL Connect Attempt.";
     
    # MySQL data source name must have a valid database as the
    # third argument; this uses the sakila demo database.
    my $dsn = "DBI:mysql:sakila";
     
    # Local variables to build the connection string.
    my $username = "student";
    my $password = "student";
     
    # Set arguments for MySQL database error management.
    my %attr = ( PrintError=>0,  # turn off error reporting via warn()
                 RaiseError=>1); # turn on error reporting via die()           
     
    # Create connction with a data source name, user name and password.
    my $dbh  = DBI->connect($dsn,$username,$password, \%attr);
     
    # Creaet a static SQL statement or query.
    my $sth = $dbh->prepare("SELECT title         \
    	                 ,      release_year  \
    			 ,      rating        \
    			 FROM   film          \
    			 WHERE  title LIKE 'roc%'");
     
    # Execute the static statement.
    $sth->execute() or die "Execution failed: $dbh->errstr()";
     
    # Read data and print by comma-delimited row position.
    print "----------------------------------------\n";
    while (my @row = $sth->fetchrow_array()) {
      print join(", ", @row), "\n";
    }
    print "----------------------------------------\n";
     
    # Close the statement.
    $sth->finish;
     
    # Disconnect from database connection.
    $dbh->disconnect();
     
    # Print with say() function valid connection message.
    say "Connected to the MySQL database.";

    After setting the permissions to -rwxr-xr-x. you call it with this CLI command:

    ./columns.pl

    It prints:

    Perl MySQL Connect Attempt.
    ----------------------------------------
    ROCK INSTINCT, 2006, G
    ROCKETEER MOTHER, 2006, PG-13
    ROCKY WAR, 2006, PG-13
    ----------------------------------------
    Connected to the MySQL database.
  5. After connecting to the database and securing variables by reference notation, lets return the value as an array of rows in a dynamic.pl program. This one just returns data from the film table of the sakila database. It is a dynamic query because a string passed to the execute method and that value is bound to a ? placeholder in the SQL statement.

    #!/usr/bin/perl
     
    # Import libraries.
    use strict;
    use warnings;
    use v5.10;     # for say() function
    use DBI;
     
    # Print with say() function message.
    say "Perl MySQL Connect Attempt.";
     
    # MySQL data source name must have a valid database as the
    # third argument; this uses the sakila demo database.
    my $dsn = "DBI:mysql:sakila";
     
    # Local variables to build the connection string.
    my $username = "student";
    my $password = "student";
     
    # Set arguments for MySQL database error management.
    my %attr = ( PrintError=>0,  # turn off error reporting via warn()
                 RaiseError=>1); # turn on error reporting via die()           
     
    # Create connction with a data source name, user name and password.
    my $dbh  = DBI->connect($dsn,$username,$password, \%attr);
     
    # Creaet a static SQL statement or query.
    my $sth = $dbh->prepare("SELECT title         \
    	                 ,      release_year  \
    			 ,      rating        \
    			 FROM   film          \
    			 WHERE  title LIKE CONCAT(?,'%')");
     
    # Execute the dynamic statement by providing an input parameter.
    $sth->execute('roc') or die "Execution failed: $dbh->errstr()";
     
    # Read data and print by comma-delimited row position.
    print "----------------------------------------\n";
    while (my @row = $sth->fetchrow_array()) {
      print join(", ", @row), "\n";
    }
    print "----------------------------------------\n";
     
    # Close the statement.
    $sth->finish;
     
    # Disconnect from database connection.
    $dbh->disconnect();
     
    # Print with say() function valid connection message.
    say "Connected to the MySQL database.";

    After setting the permissions to -rwxr-xr-x. you call it with this CLI command:

    ./dynamic.pl

    It prints:

    Perl MySQL Connect Attempt.
    ----------------------------------------
    ROCK INSTINCT, 2006, G
    ROCKETEER MOTHER, 2006, PG-13
    ROCKY WAR, 2006, PG-13
    ----------------------------------------
    Connected to the MySQL database.
  6. After connecting to the database and securing variables by reference notation, lets return the value as an array of rows in a input.pl program. This one just returns data from the film table of the sakila database. It is a dynamic query because an input parameter is passed to a local variable and the local variable is bound to a ? placeholder in the SQL statement.

    #!/usr/bin/perl
     
    # Import libraries.
    use strict;
    use warnings;
    use v5.10;     # for say() function
    use DBI;
     
    # Get the index value of the maximum argument in the
    # argument.
    my $argc = $#ARGV;
     
    # Accept first argument value as parameter.
    my $param = $ARGV[$argc];
     
    # Verify variable value assigned.
    if (not defined $param) {
      die "Need parameter value.\n";
    }
     
    # Print with say() function message.
    say "Perl MySQL Connect Attempt.";
     
    # MySQL data source name must have a valid database as the
    # third argument; this uses the sakila demo database.
    my $dsn = "DBI:mysql:sakila";
     
    # Local variables to build the connection string.
    my $username = "student";
    my $password = "student";
     
    # Set arguments for MySQL database error management.
    my %attr = ( PrintError=>0,  # turn off error reporting via warn()
                 RaiseError=>1); # turn on error reporting via die()           
     
    # Create connction with a data source name, user name and password.
    my $dbh  = DBI->connect($dsn,$username,$password, \%attr);
     
    # Creaet a static SQL statement or query.
    my $sth = $dbh->prepare("SELECT title         \
    	                 ,      release_year  \
    			 ,      rating        \
    			 FROM   film          \
    			 WHERE  title LIKE CONCAT(?,'%')");
     
    # Execute the static statement.
    $sth->execute($param) or die "Execution failed: $dbh->errstr()";
     
    # Read data and print by comma-delimited row position.
    print "----------------------------------------\n";
    while (my @row = $sth->fetchrow_array()) {
      print join(", ", @row), "\n";
    }
    print "----------------------------------------\n";
     
    # Close the statement.
    $sth->finish;
     
    # Disconnect from database connection.
    $dbh->disconnect();
     
    # Print with say() function valid connection message.
    say "Connected to the MySQL database.";

    After setting the permissions to -rwxr-xr-x. you call it with this CLI command:

    ./input.pl ta

    It prints:

    Perl MySQL Connect Attempt.
    ----------------------------------------
    TADPOLE PARK, 2006, PG
    TALENTED HOMICIDE, 2006, PG
    TARZAN VIDEOTAPE, 2006, PG-13
    TAXI KICK, 2006, PG-13
    ----------------------------------------
    Connected to the MySQL database.

I think these examples cover most of the basic elements of writing Perl against the MySQL database. If I missed something you think would be useful, please advise. As always, I hope this helps those working with the MySQL and Perl products.

Written by maclochlainn

November 17th, 2022 at 12:01 am

Oracle PLS-00103 Gotcha

without comments

Teaching PL/SQL can be fun and sometimes challenging when you need to troubleshoot a student error. Take the Oracle PLS-00103 error can be very annoying when it return like this:

24/5     PLS-00103: Encountered the symbol "LV_CURRENT_DATE" WHEN
         expecting one OF the following:
         language

Then, you look at the code and see:

22
23
24
25
   , pv_user_id             NUMBER ) IS
 
    /* Declare local constants. */
    lv_current_date      DATE := TRUNC(SYSDATE);

Obviously, there’s nothing wrong on the line number that the error message pointed. Now, here’s where it gets interesting because of a natural human failing. The student thought they had something wrong with declaring the variable and tested as stand alone procedure and anonymous block. Naturally, they were second guessing what they knew about the PL/SQL.

That’s when years of experience with PL/SQL kicks in to solve the problem. The trick is recognizing two things:

  1. The error message points to the first line of code in a package body.
  2. The error is pointing to the first character on the line after the error.

That meant that the package body was incorrectly defined. A quick check to the beginning of the package body showed:

1
2
3
4
5
6
CREATE OR REPLACE
  PACKAGE account_creation AS
 
  PROCEDURE insert_contact
  ( pv_first_name          VARCHAR2
  , pv_middle_name         VARCHAR2 := NULL

The student failed to designate the package as an implementation by omitting the keyword BODY from line 2. The proper definition of the package body should be:

1
2
3
4
5
6
CREATE OR REPLACE
  PACKAGE BODY account_creation AS
 
  PROCEDURE insert_contact
  ( pv_first_name          VARCHAR2
  , pv_middle_name         VARCHAR2 := NULL

That’s the resolution for the error message. I wrote this because I checked if they should have been able to find a helpful article with a google search. I discovered that there wasn’t an answer like this that came up after 10 minutes of various searches.

As always, I hope this helps those writing PL/SQL.

Written by maclochlainn

October 3rd, 2022 at 12:11 am

PL/SQL Overloading

without comments

So, I wrote an updated example of my grandma and tweetie_bird for my students. It demonstrates overloading with the smallest parameter lists possible across a transaction of two tables. It also shows how one version of the procedure can call another version of the procedure.

The tables are created with the following:

/* Conditionally drop grandma table and grandma_s sequence. */
BEGIN
  FOR i IN (SELECT object_name
            ,      object_type
            FROM   user_objects
            WHERE  object_name IN ('GRANDMA','GRANDMA_SEQ')) LOOP
    IF i.object_type = 'TABLE' THEN
      /* Use the cascade constraints to drop the dependent constraint. */
      EXECUTE IMMEDIATE 'DROP TABLE '||i.object_name||' CASCADE CONSTRAINTS';
    ELSE
      EXECUTE IMMEDIATE 'DROP SEQUENCE '||i.object_name;
    END IF;
  END LOOP;
END;
/
 
/* Create the table. */
CREATE TABLE GRANDMA
( grandma_id     NUMBER       CONSTRAINT grandma_nn1 NOT NULL
, grandma_house  VARCHAR2(30) CONSTRAINT grandma_nn2 NOT NULL
, created_by     NUMBER       CONSTRAINT grandma_nn3 NOT NULL
, CONSTRAINT grandma_pk       PRIMARY KEY (grandma_id)
);
 
/* Create the sequence. */
CREATE SEQUENCE grandma_seq;
 
/* Conditionally drop a table and sequence. */
BEGIN
  FOR i IN (SELECT object_name
            ,      object_type
            FROM   user_objects
            WHERE  object_name IN ('TWEETIE_BIRD','TWEETIE_BIRD_SEQ')) LOOP
    IF i.object_type = 'TABLE' THEN
      EXECUTE IMMEDIATE 'DROP TABLE '||i.object_name||' CASCADE CONSTRAINTS';
    ELSE
      EXECUTE IMMEDIATE 'DROP SEQUENCE '||i.object_name;
    END IF;
  END LOOP;
END;
/
 
/* Create the table with primary and foreign key out-of-line constraints. */
CREATE TABLE TWEETIE_BIRD
( tweetie_bird_id     NUMBER        CONSTRAINT tweetie_bird_nn1 NOT NULL
, tweetie_bird_house  VARCHAR2(30)  CONSTRAINT tweetie_bird_nn2 NOT NULL
, grandma_id          NUMBER        CONSTRAINT tweetie_bird_nn3 NOT NULL
, created_by          NUMBER        CONSTRAINT tweetie_bird_nn4 NOT NULL
, CONSTRAINT tweetie_bird_pk        PRIMARY KEY (tweetie_bird_id)
, CONSTRAINT tweetie_bird_fk        FOREIGN KEY (grandma_id)
  REFERENCES GRANDMA (GRANDMA_ID)
);
 
/* Create sequence. */
CREATE SEQUENCE tweetie_bird_seq;

The sylvester package specification holds the two overloaded procedures, like:

CREATE OR REPLACE
  PACKAGE sylvester IS
 
  /* Three variable length strings. */
  PROCEDURE warner_brother
  ( pv_grandma_house       VARCHAR2
  , pv_tweetie_bird_house  VARCHAR2
  , pv_system_user_name    VARCHAR2  );
 
  /* Two variable length strings and a number. */  
  PROCEDURE warner_brother
  ( pv_grandma_house       VARCHAR2
  , pv_tweetie_bird_house  VARCHAR2
  , pv_system_user_id      NUMBER   );
 
END sylvester;
/

The sylvester package implements two warner_brother procedures. One takes the system user’s ID and the other takes the system user’s name. The procedure that accepts the system user name queries the system_user table with the system_user_name to get the system_user_id column and then calls the other version of itself. This demonstrates how you only write logic once when overloading and let one version call the other with the added information.

Here’s the sylvester package body code:

CREATE OR REPLACE
  PACKAGE BODY sylvester IS
 
  /* Procedure warner_brother with user name. */
  PROCEDURE warner_brother
  ( pv_grandma_house       VARCHAR2
  , pv_tweetie_bird_house  VARCHAR2
  , pv_system_user_id      NUMBER  ) IS
 
    /* Declare a local variable for an existing grandma_id. */
    lv_grandma_id   NUMBER;
 
    FUNCTION get_grandma_id
    ( pv_grandma_house  VARCHAR2 ) RETURN NUMBER IS
 
      /* Initialized local return variable. */
      lv_retval  NUMBER := 0;  -- Default value is 0.
 
      /* A cursor that lookups up a grandma's ID by their name. */
      CURSOR find_grandma_id
      ( cv_grandma_house  VARCHAR2 ) IS
        SELECT grandma_id
        FROM   grandma
        WHERE  grandma_house = cv_grandma_house;
 
    BEGIN   
      /* Assign a grandma_id as the return value when a row exists. */
      FOR i IN find_grandma_id(pv_grandma_house) LOOP
        lv_retval := i.grandma_id;
      END LOOP;
 
      /* Return 0 when no row found and the grandma_id when a row is found. */
      RETURN lv_retval;
    END get_grandma_id;
 
  BEGIN
    /* Set the savepoint. */
    SAVEPOINT starting;
 
    /*
     *  Identify whether a member account exists and assign it's value
     *  to a local variable.
     */
    lv_grandma_id := get_grandma_id(pv_grandma_house);
 
    /*
     *  Conditionally insert a new member account into the member table
     *  only when a member account does not exist.
     */
    IF lv_grandma_id = 0 THEN
 
      /* Insert grandma. */
      INSERT INTO grandma
      ( grandma_id
      , grandma_house
      , created_by )
      VALUES
      ( grandma_seq.NEXTVAL
      , pv_grandma_house
      , pv_system_user_id  );
 
      /* Assign grandma_seq.currval to local variable. */
      lv_grandma_id := grandma_seq.CURRVAL;
 
    END IF;
 
    /* Insert tweetie bird. */
    INSERT INTO tweetie_bird
    ( tweetie_bird_id
    , tweetie_bird_house 
    , grandma_id
    , created_by )
    VALUES
    ( tweetie_bird_seq.NEXTVAL
    , pv_tweetie_bird_house
    , lv_grandma_id
    , pv_system_user_id );
 
    /* If the program gets here, both insert statements work. Commit it. */
    COMMIT;
 
  EXCEPTION
    /* When anything is broken do this. */
    WHEN OTHERS THEN
      /* Until any partial results. */
      ROLLBACK TO starting;
  END;
 
  PROCEDURE warner_brother
  ( pv_grandma_house       VARCHAR2
  , pv_tweetie_bird_house  VARCHAR2
  , pv_system_user_name    VARCHAR2  ) IS
 
    /* Define a local variable. */
	lv_system_user_id  NUMBER := 0;
 
    FUNCTION get_system_user_id
    ( pv_system_user_name  VARCHAR2 ) RETURN NUMBER IS
 
      /* Initialized local return variable. */
      lv_retval  NUMBER := 0;  -- Default value is 0.
 
      /* A cursor that lookups up a grandma's ID by their name. */
      CURSOR find_system_user_id
      ( cv_system_user_id  VARCHAR2 ) IS
        SELECT system_user_id
        FROM   system_user
        WHERE  system_user_name = pv_system_user_name;
 
    BEGIN   
      /* Assign a grandma_id as the return value when a row exists. */
      FOR i IN find_system_user_id(pv_system_user_name) LOOP
        lv_retval := i.system_user_id;
      END LOOP;
 
      /* Return 0 when no row found and the grandma_id when a row is found. */
      RETURN lv_retval;
    END get_system_user_id;
 
  BEGIN
 
    /* Convert a system_user_name to system_user_id. */
	lv_system_user_id := get_system_user_id(pv_system_user_name);
 
	/* Call the warner_brother procedure. */
	warner_brother
    ( pv_grandma_house      => pv_grandma_house
    , pv_tweetie_bird_house => pv_tweetie_bird_house
    , pv_system_user_id     => lv_system_user_id  );
 
  EXCEPTION
    /* When anything is broken do this. */
    WHEN OTHERS THEN
      /* Until any partial results. */
      ROLLBACK TO starting;
  END;  
 
END sylvester;
/

The following anonymous block test case works with the code:

BEGIN
  sylvester.warner_brother( pv_grandma_house      => 'Blue House'
                          , pv_tweetie_bird_house => 'Cage'
				          , pv_system_user_name   => 'DBA 3' );
  sylvester.warner_brother( pv_grandma_house      => 'Blue House'
                          , pv_tweetie_bird_house => 'Tree House'
				          , pv_system_user_id     =>  4 );
END;
/

You can now query the results with this SQL*PLus formatting and query:

/* Query results from warner_brother procedure. */
COL grandma_id          FORMAT 9999999  HEADING "Grandma|ID #"
COL grandma_house       FORMAT A14      HEADING "Grandma House"
COL created_by          FORMAT 9999999  HEADING "Created|By"
COL tweetie_bird_id     FORMAT 9999999  HEADING "Tweetie|Bird ID"
COL tweetie_bird_house  FORMAT A18      HEADING "Tweetie Bird House"
SELECT *
FROM   grandma g INNER JOIN tweetie_bird tb
ON     g.grandma_id = tb.grandma_id;

You should see the following data:

 Grandma                 Created  Tweetie                     Grandma  Created
    ID # Grandma House        By  Bird ID Tweetie Bird House     ID #       By
-------- -------------- -------- -------- ------------------ -------- --------
       1 Blue House            3        1 Cage                      1        3
       1 Blue House            3        2 Tree House                1        4

As always, I hope complete code samples help solve real problems.

Written by maclochlainn

September 29th, 2022 at 9:24 pm

PostgreSQL Trigger 1

without comments

This entry covers how to write a statement logging trigger for PostgreSQL. It creates two tables: avenger and avenger_log; one avenger_t1 trigger, and a testing INSERT statement.

It was written to help newbies know how and what to return from a function written for a statement-level trigger. They often get stuck on the following when they try to return true. The term non-composite is another way to describe the tuple inserted.

psql:basics_postgres.sql:
59: ERROR:  cannot return non-composite value from function returning composite type
CONTEXT:  PL/pgSQL function write_avenger_t1() line 15 at RETURN

The avenger table:

/* Conditionally drop table. */
DROP TABLE IF EXISTS avenger;
 
/* Create table. */
CREATE TABLE avenger
( avenger_id    SERIAL
, avenger_name  VARCHAR(30)
, first_name    VARCHAR(20)
, last_name     VARCHAR(20));

Seed the avenger table:

/* Seed the avenger table with data. */
INSERT INTO avenger
( first_name, last_name, avenger_name )
VALUES
 ('Anthony', 'Stark', 'Iron Man')
,('Thor', 'Odinson', 'God of Thunder')
,('Steven', 'Rogers', 'Captain America')
,('Bruce', 'Banner', 'Hulk')
,('Clinton', 'Barton', 'Hawkeye')
,('Natasha', 'Romanoff', 'Black Widow')
,('Peter', 'Parker', 'Spiderman')
,('Steven', 'Strange', 'Dr. Strange')
,('Scott', 'Lange', 'Ant-man');

The avenger_log table:

/* Conditionally drop table. */
DROP TABLE IF EXISTS avenger_log;
 
/* Create table. */
CREATE TABLE avenger_log
( avenger_log_id  SERIAL
, trigger_name    VARCHAR(30)
, trigger_timing  VARCHAR(6)
, trigger_event   VARCHAR(6)
, trigger_type    VARCHAR(12));

The INSERT statement that tests the trigger:

DROP FUNCTION IF EXISTS avenger_t1_function; 
CREATE FUNCTION avenger_t1_function()
  RETURNS TRIGGER AS
$$
BEGIN
  /* Insert a row into the avenger_log table.
   * Also, see PostrgreSQL 39.9 Trigger Procedures. */
  INSERT INTO avenger_log
  ( trigger_name
  , trigger_timing
  , trigger_event
  , trigger_type )
  VALUES
  ( UPPER(TG_NAME)
  , TG_WHEN
  , TG_OP
  , TG_LEVEL );
 
  /* A statement trigger doesn't use a composite type or tuple,
   * it should simply return an empty composite type or void. */
  RETURN NULL;
END;
$$ LANGUAGE plpgsql;

The avenger_t1 statement trigger:

CREATE TRIGGER avenger_t1
  BEFORE INSERT ON avenger
  EXECUTE FUNCTION avenger_t1_function();

The INSERT statement:

INSERT INTO avenger
( first_name, last_name, avenger_name )
VALUES
 ('Hope', 'van Dyne', 'Wasp');

The results logged to the avenger_log table from a query:

 avenger_log_id | trigger_name | trigger_timing | trigger_event | trigger_type
----------------+--------------+----------------+---------------+--------------
              1 | AVENGER_T1   | BEFORE         | INSERT        | STATEMENT
(1 row)

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

Written by maclochlainn

July 11th, 2022 at 3:12 pm

PL/pgSQL Transactions

without comments

There are many nuances that I show students about PL/pgSQL because first I teach them how to use PL/SQL. These are some of the differences:

  • PL/SQL declares the function or procedure and then uses the IS keyword; whereas, PL/pgSQL uses the AS keyword.
  • PL/SQL uses the RETURN keyword for functions declarations, like:

    RETURN [data_type} IS

    Whereas, PL/pgSQL uses the plural RETURNS keyword in the function declaration, like:

    RETURNS [data_type] AS
  • PL/SQL considers everything after the function or procedure header as the implicit declaration section; whereas, PL/pgSQL requires you block the code with something like $$ (double dollar symbols) and explicitly use the DECLARE keyword.
  • PL/SQL supports local functions (inside the DECLARE block of a function or procedure); whereas, PL/pgSQL doesn’t.
  • PL/SQL puts the variable modes (IN, INOUT, OUT) between the parameter name and type; whereas, PL/pgSQL puts them before the variable name.
  • PL/SQL declares cursors like:

    CURSOR cursor_name (parameter_list) IS

    Whereas, PL/pgSQL declares them like

    cursor_name CURSOR (parameter_list) FOR
  • PL/SQL terminates and runs the block by using an END keyword, an optional module name, a semicolon to terminate the END; statement, and a forward slash to dispatch the program to PL/SQL statement engine:

    END [module_name];
    /

    Whereas, PL/pgSQL terminates and runs the block by using an END keyword, a semicolon to terminate the END; statement, two dollar signs to end the PL/pgSQL block, and a semicolon to dispatch the program.

    END LANGUAGE plpgsql;
    $$;

After all that basic syntax discussion, we try to create a sample set of tables, a function, a procedure, and a test case in PL/pgSQL. They’ve already done a virtually equivalent set of tasks in PL/SQL.

Here are the steps:

  1. Create the grandma and tweetie_bird tables:

    /* Conditionally drop grandma table and grandma_s sequence. */
    DROP TABLE IF EXISTS grandma CASCADE;
     
    /* Create the table. */
    CREATE TABLE GRANDMA
    ( grandma_id     SERIAL
    , grandma_house  VARCHAR(30)  NOT NULL
    , PRIMARY KEY (grandma_id)
    );
     
    /* Conditionally drop a table and sequence. */
    DROP TABLE IF EXISTS tweetie_bird CASCADE;
     
    /* Create the table with primary and foreign key out-of-line constraints. */
    SELECT 'CREATE TABLE tweetie_bird' AS command;
    CREATE TABLE TWEETIE_BIRD
    ( tweetie_bird_id     SERIAL
    , tweetie_bird_house  VARCHAR(30)   NOT NULL
    , grandma_id          INTEGER       NOT NULL
    , PRIMARY KEY (tweetie_bird_id)
    , CONSTRAINT tweetie_bird_fk        FOREIGN KEY (grandma_id)
      REFERENCES grandma (grandma_id)
    );
  2. Create a get_grandma_id function that returns a number, which should be a valid primary key value from the grandma_id column of the grandma table.

    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
    
    CREATE OR REPLACE
      FUNCTION get_grandma_id
      ( IN pv_grandma_house  VARCHAR ) RETURNS INTEGER AS
    $$
      /* Required for PL/pgSQL programs. */
      DECLARE
     
        /* Local return variable. */
        lv_retval  INTEGER := 0;  -- Default value is 0.
     
        /* Use a cursor, which will not raise an exception at runtime. */
        find_grandma_id CURSOR 
        ( cv_grandma_house  VARCHAR ) FOR
          SELECT grandma_id
          FROM   grandma
          WHERE  grandma_house = cv_grandma_house;
     
      BEGIN  
     
        /* Assign a value when a row exists. */
        FOR i IN find_grandma_id(pv_grandma_house) LOOP
          lv_retval := i.grandma_id;
        END LOOP;
     
        /* Return 0 when no row found and the ID # when row found. */
        RETURN lv_retval;
      END;
    $$ LANGUAGE plpgsql;
  3. Create a Warner_brother procedure that writes data across two tables as a transaction. You con’t include any of the following in your functions or procedures because all PostgreSQL PL/pgSQL functions and procedures are transaction by default:

    • SET TRANSACTION
    • START TRANSACTION
    • SAVEPOINT
    • COMMIT

    A ROLLBACK should be placed in your exception handler as qualified on lines #33 thru #36. The warner_brother procedure inserts rows into the grandma and tweetie_bird tables.

    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
    
    /* Create or replace procedure warner_brother. */
    CREATE OR REPLACE
      PROCEDURE warner_brother
      ( pv_grandma_house       VARCHAR
      , pv_tweetie_bird_house  VARCHAR ) AS
    $$ 
      /* Required for PL/pgSQL programs. */
      DECLARE
     
      /* Declare a local variable for an existing grandma_id. */
      lv_grandma_id   INTEGER;
     
    BEGIN  
      /* Check for existing grandma row. */
      lv_grandma_id := get_grandma_id(pv_grandma_house);
      IF lv_grandma_id = 0 THEN 
        /* Insert grandma. */
        INSERT INTO grandma
        ( grandma_house )
        VALUES
        ( pv_grandma_house )
        RETURNING grandma_id INTO lv_grandma_id;
      END IF;
     
      /* Insert tweetie bird. */
      INSERT INTO tweetie_bird
      ( tweetie_bird_house 
      , grandma_id )
      VALUES
      ( pv_tweetie_bird_house
      , lv_grandma_id );
     
    EXCEPTION
      WHEN OTHERS THEN
        ROLLBACK;
        RAISE NOTICE '[%] [%]', SQLERRM, SQLSTATE;  
    END;
    $$ LANGUAGE plpgsql;

    You should take note of the RETURNING-INTO statement on line #22. The alternative to this clause isn’t pretty if you know that PostgreSQL uses a table name, column name, and the literal seq value separated by underscores (that is, snake case), like:

        /* Assign current value to local variable. */
        lv_grandma_id := CURRVAL('grandma_grandma_id_seq');

    It would be even uglier if you had to look up the sequence name, like:

        /* Assign current value to local variable. */
        lv_grandma_id := CURRVAL(pg_get_serial_sequence('grandma','grandma_id'));
  4. You can test the combination of these two stored procedures with the following DO-block:

    /* Test the warner_brother procedure. */
    DO
    $$
    BEGIN
      /* Insert the yellow house. */
      CALL warner_brother( 'Yellow House', 'Cage');
      CALL warner_brother( 'Yellow House', 'Tree House');
     
      /* Insert the red house. */
      CALL warner_brother( 'Red House', 'Cage');
      CALL warner_brother( 'Red House', 'Tree House');
    END;
    $$ LANGUAGE plpgsql;

    Then, query the results:

    SELECT *
    FROM   grandma g INNER JOIN tweetie_bird tb
    ON.    g.grandma_id = tb.grandma_id;

    It should return:

     grandma_id | grandma_house | tweetie_bird_id | tweetie_bird_house | grandma_id
    ------------+---------------+-----------------+--------------------+------------
              1 | Red House     |               1 | Cage               |          1
              1 | Red House     |               2 | Tree House         |          1
              2 | Yellow House  |               3 | Cage               |          2
              2 | Yellow House  |               4 | Tree House         |          2
    (4 rows)
  5. As always, I hope writing a clear and simple examples helps those looking for sample code.

Written by maclochlainn

June 16th, 2022 at 9:38 pm

MySQL Partitioned Tables

with one comment

MySQL Partitioned Tables

Learning Outcomes

  • Learn about List Partitioning.
  • Learn about Range Partitioning.
  • Learn about Columns Partitioning.
  • Learn about Hash Partitioning.
  • Learn about Key Partitioning.
  • Learn about Subpartitioning.

Lesson Material

MySQL supports partitioning of tables. It supports range, list, hash, and key partitioning. Range partitioning lets you partition based on column values that fall within given ranges. List partitioning lets you partition based on columns matching one of a set of discrete values. Hash partitioning lets you partition based on the return value from a user-defined expression (the result from a stored SQL/PSM function). Key partitioning performs like hash partitioning, but it lets a user select one or more columns from the set of columns in a table; a hash manages the selection process for you. A hash is a method of organizing keys to types of data, and hashes speed access to read and change data in tables.

Each of the following subsections discusses one of the supported forms of partitioning in MySQL. Naturally, there are differences between other databases and MySQL’s implementation.

List Partitioning

A MySQL list partition works by identifying a column that contains an integer value, the franchise_number in the following example. Partitioning clauses follow the list of columns and constraints and require a partitioning key to be in the primary key or indexed.

The following list partition works with literal numeric values. MySQL uses the IN keyword for list partitions. Note that there’s no primary key designated and an index is on the auto-incrementing surrogate key column. A complete example is provided to avoid confusion on how to index the partitioning key:

CREATE TABLE franchise
( franchise_id     INT UNSIGNED AUTO_INCREMENT
, franchise_number INT UNSIGNED
, franchise_name   VARCHAR(20)
, city             VARCHAR(20)
, state            VARCHAR(20)
, index idx (franchise_id))
PARTITION BY LIST(franchise_number)
( PARTITION offshore VALUES IN (49,50)
, PARTITION west VALUES IN (34,45,48)
, PARTITION desert VALUES IN (46,47)
, PARTITION rockies VALUES IN (38,41,42,44));

The inclusion of a PRIMARY KEY constraint on the franchise_id column would trigger an ERROR 1503 when the partitioning key isn’t the primary key. The reason for the error message is that a primary key implicitly creates a unique index, and that index would conflict with the partitioning by list instruction. The use of a non-unique idx index on the franchise_id column is required when you want to partition on a non-primary key column.

Range Partitioning

Range partitioning works only with an integer value or an expression that resolves to an integer against the primary key column. The limitation of the integer drives the necessity of choosing an integer column for range partitioning. You can’t define a range-partitioned table with a PRIMARY KEY constraint unless the primary key becomes your partitioning key, like
the one below.

CREATE TABLE ordering
( ordering_id     INT UNSIGNED AUTO_INCREMENT
, item_id         INT UNSIGNED
, rental_amount   DECIMAL(15,2)
, rental_date     DATE
, index idx (ordering_id))
PARTITION BY RANGE(item_id)
( PARTITION jan2011 VALUES LESS THAN (10000) , PARTITION feb2011 VALUES LESS THAN (20000) , PARTITION mar2011 VALUES LESS THAN (30000));

Range partitioning is best suited to large tables that you want to break into smaller pieces based on the integer column. You can also use stored functions that return integers as the partitioning key instead of the numeric literals shown. Few other options are available in MySQL.

Columns Partitioning

Columns partitioning is a new variant of range and list partitioning. It is included in MySQL 5.5 and forward. Both range and list partitioning work on an integer-based column (using TINYINT, SMALLINT, MEDIUMINT, INT [alias INTEGER], and BIGINT). Columns partitioning extends those models by expanding the possible data types for the partitioning column to include CHAR, VARCHAR, BINARY, and VARBINARY string data types, and DATE, DATETIME, or TIMESTAMP data types. You still can’t use other number data types such as DECIMAL and FLOAT. The TIMESTAMP data type is also available only in range partitions with the caveat that you use a UNIX_TIMESTAMP function, according to MySQL Bug 42849.

Hash Partitioning

Hash partitions ensure an even distribution of rows across a predetermined number of partitions. It is probably the easiest way to partition a table quickly to test the result of partitioning on a large table. You should base hash partitions on a surrogate or natural primary key.

The following provides a modified example of the ordering table:

CREATE TABLE ordering
( ordering_id       INT UNSIGNED PRIMARY KEY AUTO_INCREMENT
, item_id           INT UNSIGNED
, rental_amount     DECIMAL(15,2)
, rental_date       DATE)
PARTITION BY HASH(ordering_id) PARTITIONS 8;

This is the partitioning type that benefits from a PRIMARY KEY constraint because it automatically creates a unique index that can be used by the hash. A non-unique index such as the list partitioning example doesn’t work for a hash partition.

Key Partitioning

Key partitioning is valuable because you can partition on columns that aren’t integers. It performs along the line of hash partitioning, except the MySQL Server uses its own hashing expression.

CREATE TABLE orders_list
( order_list_id     INT UNSIGNED AUTO_INCREMENT
, customer_surname  VARCHAR(30)
, store_id          INT UNSIGNED
, salesperson_id    INT UNSIGNED
, order_date        DATE
, index idx (order_list_id))
PARTITION BY KEY (order_date) PARTITIONS 8;

This is the only alternative when you want to partition by date ranges. Like the hash partition, it’s easy to deploy. The only consideration is the number of slices that you want to make of the data in the table.

Subpartitioning

The concept of subpartitioning is also known as composite partitioning. You can subpartition range or list partitions with a hash, linear hash, or linear key.

A slight change to the previously created ordering table is required to demonstrate composite partitioning: we’ll add a store_id column to the table definition. The following is an example of a range partition subpartitioned by a hash:

CREATE TABLE ordering
INT UNSIGNED
INT UNSIGNED
DATE
 ( ordering_id     INT UNSIGNED AUTO_INCREMENT
, item_id          INT UNSIGNED
, store_id         INT UNSIGNED
, rental_amount    DECIMAL(15,2)
, rental_date      DATE
, index idx (ordering_id))
PARTITION BY RANGE(item_id)
  SUBPARTITION BY HASH(store_id) SUBPARTITIONS 4
( PARTITION jan2011 VALUES LESS THAN (10000)
, PARTITION feb2011 VALUES LESS THAN (20000)
, PARTITION mar2011 VALUES LESS THAN (30000));

Composite partitioning is non-trivial and might require some experimentation to achieve optimal results. Plan on making a few tests of different scenarios before you deploy a solution.

Written by maclochlainn

May 30th, 2022 at 10:08 pm