Database Settings

备注

For an introduction of internal data storage, please refer to Introduction of digiKam Databases section.

The Sqlite Database

SQLite is a relational database management system, written in C programming library. SQLite is not directly comparable to client/server SQL database engines such as MySQL, Oracle or PostgreSQL. Rather, it is an embedded SQL database engine, i.e. it is embedded in an end program. SQLite reads and writes directly to ordinary disk files. For device-local storage with low writer concurrency and less than a terabyte of content, SQLite is almost always a better solution. SQLite is fast and reliable and it requires no configuration or maintenance. It keeps things simple. SQLite “just works”.

By default, digiKam uses SQLite as its back-end for storing important metadata and thumbnails. Three SQLite files used for storing them are named respectively:

Database

File-Name

Core

digikam4.db

Thumbs

thumbnails-digikam.db

Similarity

similarity.db

Faces

recognition.db

To make your application run fast and smoothly, it is recommended to check and optimize your databases once in awhile. This could be achieved with the menu option Tools ‣ Maintenance… and the stage Perform Database Cleaning. See this Maintenance tool section for details. A recommended tool is SQLite Browser, a high quality and easy to use visual tool for managing database objects. For Ubuntu and its derivatives, it could be retrieved using sudo apt install sqlitebrowser. Now you can switch to the directory where databases are stored and visualize the database contents.

备注

Take care to use a place hosted by fast hardware (such as SSD or NVMe) with enough free space especially for thumbnails database. A remote file system such as NFS cannot be used here. For performance and technical reasons relevant of SQLite, you cannot use a media from the network.

SQLite database files could be found in your collection folder, which you have added to digiKam. (By default, if you add your “Pictures” collection, the database files will be present in ~/Pictures folder).

The digiKam SQLite Configuration Page

The WAL SQLite mode is a very important option that we recommend to turn on with large databases to optimize transactions and improve performances.

The MySQL Database

MySQL Versus SQLite

MySQL is an open-source, relational database management system, written in C and C++. Original development of MySQL by Michael Widenius and David Axmark beginning in 1994. Sun Microsystems acquired MySQL in 2008, which was later acquired by Oracle in 2010. MySQL currently works on almost all system platforms (Linux, Microsoft Windows, OS X, SunOS …).

MariaDB server is a community developed fork of MySQL server. Started by core members of the original MySQL team, MariaDB actively works with outside developers to deliver the most featureful, stable, and sanely licensed open SQL server in the industry.

MariaDB has actually overtaken MySQL, because of few basic reasons:

  • MariaDB development is more open and vibrant.

  • More cutting edge features.

  • More storage engines.

  • Better performance.

  • Compatible and easy to migrate.

digiKam also provides support for popular MySQL database engine. Of course, you might wonder why you’d want to switch to MySQL when SQLite already does a good job of managing the data? MySQL offers many advantages for storing digiKam data, especially when collections include more than 100,000 items. With such large collections, SQLite introduces latency which slows down the application.

备注

With WAL option enabled, SQLite can be easily used for more than 100,000 items especially with an SSD or NVMe storage. It must be even faster than MySQL and more stable. See this page for technical details.

Using MySQL as digiKam’s database back-end allows you to store the data on local as well as remote server. Local, to replace the local SQLite storage and latter, to use a shared computer through network. Using MySQL as digiKam’s database back-end allows you to store the data on a remote server. This way, you can use multiple digiKam installations (For instance,on your notebook and PC) to access and manage your photo collections. You can also use MySQL tools to backup and analyze digiKam’s data.

To switch from SQLite to MySQL database, go to Settings ‣ Configure digiKam… and then under Database section, select a database from the drop down list.

  • MySQL Internal: This allows to run an internal database server on your system. digiKam uses Unix socket for the connection.

  • MySQL Server: Use this if you’ve your data on remote server and you’re on a different machine trying to access the collection.

The MySQL Internal Server

While using a large collection hosted on hard drive (HDD - not SSD or NVMe device), with a size greater than 100,000 items, the application tends to slow down. To avoid the delay and maintain efficiency, digiKam provides option of using MySQL Internal. To be clear, this isn’t an actual server, or a public network. Instead, it is a server that runs only while application is running.

Internal server creates a separate database that can be accessed (only while application is running) using the command:

mysql --socket=/home/[user_name]/.local/share/digikam/db_misc/mysql.socket digikam

Internal server uses tree MySQL Binary Tools: mysql_install_db, mysqladmin, and mysqld. You can configure their locations in the configuration dialog. digiKam will try to find these binaries automatically if they’re installed on your system.

The digiKam MySQL Internal Configuration Page

The MySQL Remote Server

Obviously, to use digiKam with a remote MySQL, you would require a MySQL server. Or, you could also install MariaDB, which serves the purpose well. (Could be installed easily using this link.)

Follow the instructions below, if you don’t have a dedicated user account and a digiKam database already set up. Run the commands in MySQL server, after replacing password with correct one.

备注

You can select any database name. (Here it is, digikam). Just remember to fill in the database name correctly in Core, Thumbs, Similarity, and Face database names from the dialog box shown below.

CREATE USER ''@'%' IDENTIFIED BY 'password';
GRANT ALL ON *.* TO ''@'%' IDENTIFIED BY 'password';
CREATE DATABASE digikam;
GRANT ALL PRIVILEGES ON digikam.* TO ''@'%';
FLUSH PRIVILEGES;

备注

If you have an enormous collection, it’s recommended to start the MySQL server with mysql –max_allowed_packet = 128M

小技巧

If you have problems with a MySQL server on Ubuntu based Linux system, use the addition command in the mysql prompt to be able to create MySQL triggers.

SET global log_bin_trust_function_creators=1;

Now, in digiKam, go to Settings ‣ Configure digiKam… and then under Database section, select MySQL Server from the drop down list.

The digiKam Remote Mysql Configuration Page

Enter the IP address of your MySQL server in the Host Name field and specify the correct port in the Host Port field (the default port is 3306).

In the Core Db Name field, enter the name of the first database for storing photo metadata.

Specify the name of the second database for storing wavelets compressed thumbnails in the Thumbs Db Name field.

The third database is dedicated to store the similarity finger-prints performed by the fuzzy search engine. Use the Similarity Db Name field for that.

The last database is dedicated to store face histograms for recognition purpose. Use the Face Db Name field for that.

To be connected safety to the remote server, enter your MySQL identification using User and Password fields.

To check whether the database connection works properly, press the Check Connection button. If everything works as it’s supposed to, switch to the Collections sections, and add the directories containing your photos. Hit OK, and wait till digiKam populates the databases with data from photos. This can take a while if you have a lot of items to register in database.

There are some tips and recommendation to obtain the best results with a remote MySQL database server.

With slow network, digiKam hangs a lot of time especially when album contains many items (>1,000). This solution relies on network performances. Problem has been reproducible using Wifi connection, for instance. Switching to Ethernet must solves the problem.

Also, if you have an enormous collection, you should start the MySQL server with mysql –max_allowed_packet = 128M. If you’re well acquainted with using MySQL, you could also change your settings in my.ini or ~/.my.cnf files.

警告

The locale used in the Mysql server must be the same than the locale from the computer used to run digiKam to prevent problem with the double values saved in the tables of the database.

Database Type Criteria

See the resume below to choose the right database type depending of the use-cases.

Storage

Type

Items

Remarks

HDD

SQLite

< 100K

Warning: WAL is mandatory.

HDD

MySQL-Internal

> 100K

SSD

SQLite

WAL is optional.

SSD

MySQL-Internal

NVME

SQLite

WAL is optional.

NVME

MySQL-Internal

Removable

SQLite

< 100K

Warning: WAL is mandatory. USB 3.1 minimum with NVMe drive.

Removable

MySQL-Internal

> 100K

Warning: USB 3.1 minimum with NVMe drive.

Network FS

SQLite

Prohibited: SQLite databases must be stored on local file system.

Network FS

MySQL-Internal

Prohibited: MySQL databases must be stored on local file system.

Remote

MySQL-Server

MariaDB server is supported. Gigabit Ethernet or higher is recommended.
HDD

Hard Disk Drive.

SSD

Solid State Drive.

NVMe

Non-Volatile Memory.

Removable

External USB HDD/SSD/NVMe drive.

Network FS

Network File System mounted locally.

Remote

Network server as NAS (Network Attached Storage).

WAL

Write-Ahead Lock (SQLite database only).

备注

See this Digital Asset Management chapter for more details about media and data protection.

See also this Collection Settings chapter for more details about the way to configure your collections depending of your storage policy.

重要

If you share the same Removable media to host databases and/or collections between different computers, you must have the same kind of operating system, the same mount paths everywhere (use symbolic links to revolve paths), and the same digiKam version everywhere to prevent conflicts with database schemes.

If you use a common Remote server to host databases and collections, you must use the same digiKam version everywhere to prevent conflicts with database schemes. Computers running digiKam cannot be used at the same time on collections.

If you use a common Remote server to host collections, as databases are located on computers, different versions of digiKam can be used and digiKam sessions can run at the same time on collections. Take a care about concurrency access on files metadata if you turned on this option on Metadata Setup Page.

Database Migration

The photo management application comes up with an exclusive tool named Database Migration, that allows users to migrate their data. Suppose, you’re using SQLite and you wish to move all data to MySQL database, migration tool will help you do so. It can help you migrate data from SQLite to MySQL and vice versa.

To migrate to another database, go to Settings ‣ Database Migration…. A dialog box appears:

The digiKam Database Migration Tool

Now choose appropriate database types you want to convert to. Finally, click on Migrate button to convert the database from SQLite to MySQL (or vice versa). Depending of the database size this operation can take a while.

备注

Only the digiKam Core database will be migrated while conversion process. All other databases needs to be rebuilt as post-processing with the Maintenance Tools. The Thumbs and Similarity databases needs to be created from scratch, and the Face database needs the option Rebuild the Training Data.

Database Backup Recommendation

For security reasons, planing a database backup using crontab over the network can help against device dysfunctions. A NAS or an external drive can also be used for that.

Each database can be named with a different name, not only digikam. This allows to users to backup only what is needed. For instance, naming Core database as digiKam_Core, allows to isolate only this table (the most important file). Thumbnails, Similarity and Face Recognition databases can always be regenerated for scratch.

The chapter about digiKam Maintenance Tools will explain how to maintain in time the database contents and how to synchronize the collections with databases information (and vice versa).

Database Statistics

digiKam provides a unique tool to show the statistics from your collections. It includes count of images, videos (including individual count by image format), tags etc. Also, includes the Database Backend (QSQLITE or QMYSQL) and the Database Path (where your collection is located).

You can view your statistics by going to Help ‣ Database Statistics…. A dialog box like this will appear:

The digiKam Database Statistics Dialog