Difference between revisions of "MusicBrainz Database/Schema"

From MusicBrainz Wiki
Jump to navigationJump to search
(added cdstubs)
(→‎Musical work inquiring: Add diagram and rephrase after it)
 
(73 intermediate revisions by 7 users not shown)
Line 1: Line 1:
  +
{{Work in progress}}
  +
  +
This page is intended for developers querying the MusicBrainz database directly through '''PostgreSQL'''.
  +
  +
== Overview ==
  +
  +
The SQL scripts that create the schema can be found in our [https://github.com/metabrainz/musicbrainz-server/tree/master/admin/sql source code repository].
  +
  +
The database is structured around primary [[Entity|entities]] which can be edited, searched for, referred to by MBID, and linked to each other through [[Relationship|relationships]] or foreign key constraints in some case.
  +
Secondary entities cannot be linked through relationships, instead they are directly linked through foreign key constraints only.
  +
Each primary entity type has a main table sometimes accompanied by complementary tables of which a few are unique but most are common.
  +
  +
* The first diagram below shows the main tables for primary entity types and unique tables connecting these tables, that is, beyond relationships. The main tables are highlighted. All tables are shortened to focus on foreign key constraints.
  +
* The second diagram additionally shows the main tables for primary entity types and tables unique to some of these, that is, beyond aliases, annotations, edits, redirects, relationships, ratings, tags.
  +
* The third diagram shows all the tables but is too dense to just distinguish foreign key constraints for example, thus the many following sections.
  +
  +
[[Image:entity_network_overview.svg|x600px|class=zoomable|Overview of tables connecting core entities]]
  +
[[Image:entity_network_details.svg|x600px|class=zoomable|Tables specifically related to core entities]]
  +
[[Image:soup.svg|x600px|class=zoomable|All tables (exhaustive but indigestible)]]
  +
 
== Primary entities ==
 
== Primary entities ==
 
Each primary entity has a main table with the same name as the entity, containing its basic data.
 
Each primary entity has a main table with the same name as the entity, containing its basic data.
Line 4: Line 24:
 
=== [[Area]] ===
 
=== [[Area]] ===
   
  +
{|
A country, region, city or the like.
 
  +
|[[Image:area_entity_details.svg|60px|class=zoomable|Tables for the “area” core entity type properties]]
  +
|A country, region, city or the like.
   
Areas that can be used for filling in the ''Release country'' field of releases are listed, by ID, in the '''country_area''' table.
+
Areas that can be used for filling in the ''Release country'' field of releases are listed, by ID, in the <code>country_area</code> table.
  +
  +
The table <code>area_containment</code> is [https://github.com/metabrainz/musicbrainz-server/blob/88e3d5ce55179aa23e9deccf171e4e1a77efe46e/INSTALL.md#build-materialized-tables materialized (m)].
  +
|}
   
 
=== [[Artist]] ===
 
=== [[Artist]] ===
   
  +
{|
An artist is generally a musician, a group of musicians, or another music professional (composer, engineer, illustrator, producer, etc.)
 
  +
|[[Image:artist_entity_details.svg|60px|class=zoomable|Tables for the “artist” core entity type properties]]
  +
|An artist is generally a musician, a group of musicians, or another music professional (composer, engineer, illustrator, producer, etc.)
  +
  +
The tables <code>artist_release</code> and <code>artist_release_group</code> are [https://github.com/metabrainz/musicbrainz-server/blob/88e3d5ce55179aa23e9deccf171e4e1a77efe46e/INSTALL.md#build-materialized-tables materialized (m)].
  +
|}
  +
  +
=== [[Event]] ===
  +
  +
{|
  +
|[[Image:event_entity_details.svg|60px|class=zoomable|Tables for the “event” core entity type properties]]
  +
|An event refers to an organised event which people can attend, and is relevant to MusicBrainz. Generally this means live performances, like concerts and festivals.
  +
|}
  +
  +
=== [[Genre]] ===
  +
  +
{|
  +
|[[Image:genre_entity_details.svg|60px|class=zoomable|Tables for the “genre” core entity type properties]]
  +
|A genre is a descriptor for the style and conventions followed by a piece of music.
  +
|}
  +
  +
=== [[Instrument]] ===
  +
  +
{|
  +
|[[Image:instrument_entity_details.svg|60px|class=zoomable|Tables for the “instrument” core entity type properties]]
  +
|Instruments are devices created or adapted to make musical sounds. We also list common instrument groupings (such as string quartet) as instruments.
  +
|}
   
 
=== [[Label]] ===
 
=== [[Label]] ===
   
  +
{|
Labels represent mostly (but not only) imprints.
 
  +
|[[Image:label_entity_details.svg|60px|class=zoomable|Tables for the “label” core entity type properties]]
  +
|Labels represent mostly (but not only) imprints.
  +
|}
   
 
=== [[Place]] ===
 
=== [[Place]] ===
   
  +
{|
A venue, studio or other place where music is performed, recorded, engineered, etc.
 
  +
|[[Image:place_entity_details.svg|60px|class=zoomable|Tables for the “place” core entity type properties]]
  +
|A venue, studio or other place where music is performed, recorded, engineered, etc.
  +
|}
   
 
=== [[Recording]] ===
 
=== [[Recording]] ===
   
  +
{|
Represents a unique mix or edit. Has title, artist credit, duration, list of [[PUID]]s and [[ISRC]]s. Examples (all are different Recordings):
 
  +
|[[Image:recording_entity_details.svg|60px|class=zoomable|Tables for the “recording” core entity type properties]]
  +
|Represents a unique mix or edit. Has title, artist credit, duration, list of [[ISRC]]s. Examples (all are different Recordings):
   
 
* Album version of the track "''Into the Blue''" by "''Moby''"
 
* Album version of the track "''Into the Blue''" by "''Moby''"
 
* Remix "''Into the Blue (Buzz Boys Main Room Mayhem mix)''" by "''Moby''"
 
* Remix "''Into the Blue (Buzz Boys Main Room Mayhem mix)''" by "''Moby''"
 
* Remix "''Into the Blue (Underground mix)''" by "''Moby''"
 
* Remix "''Into the Blue (Underground mix)''" by "''Moby''"
  +
  +
The table <code>recording_first_release_date</code> is [https://github.com/metabrainz/musicbrainz-server/blob/88e3d5ce55179aa23e9deccf171e4e1a77efe46e/INSTALL.md#build-materialized-tables materialized (m)].
  +
|}
   
 
=== [[Release]] ===
 
=== [[Release]] ===
   
  +
{|
Real-world release object you can buy in your music store. It has release date and country, list of catalog number and label pairs, [[Release/Packaging|packaging type]] and release status. Examples:
 
  +
|[[Image:release_entity_details.svg|60px|class=zoomable|Tables for the “release” core entity type properties]]
  +
|Real-world release object you can buy in your music store. It has release date and country, list of catalog number and label pairs, [[Release/Packaging|packaging type]] and release status. Examples:
   
 
* 1984 US release of "The Wall" by "[[Artist:83d91898-7763-47d7-b03b-b92132375c47|Pink Floyd]]", release on label "Columbia Records" with catalog number "C2K 36183" and UPC "074643618328", it's an official release and comes with two CDs in jewel case.
 
* 1984 US release of "The Wall" by "[[Artist:83d91898-7763-47d7-b03b-b92132375c47|Pink Floyd]]", release on label "Columbia Records" with catalog number "C2K 36183" and UPC "074643618328", it's an official release and comes with two CDs in jewel case.
  +
  +
The tables <code>artist_release</code> and <code>release_first_release_date</code> are [https://github.com/metabrainz/musicbrainz-server/blob/88e3d5ce55179aa23e9deccf171e4e1a77efe46e/INSTALL.md#build-materialized-tables materialized (m)].
  +
|}
   
 
=== [[Release Group|Release group]] ===
 
=== [[Release Group|Release group]] ===
   
  +
{|
Represents an abstract "album" (or "single", or "EP") entity. Technically it's a group of releases, with a specified type. Examples:
 
  +
|[[Image:release_group_entity_details.svg|60px|class=zoomable|Tables for the “release group” core entity type properties]]
  +
|Represents an abstract "album" (or "single", or "EP") entity. Technically it's a group of releases, with a specified type. Examples:
   
 
* Single "Under Pressure" by "[[Artist:0383dadf-2a4e-4d10-a46a-e9e041da8eb3|Queen]] & [[Artist:5441c29d-3602-4898-b1a1-b77fa23b8e50|David Bowie]]"
 
* Single "Under Pressure" by "[[Artist:0383dadf-2a4e-4d10-a46a-e9e041da8eb3|Queen]] & [[Artist:5441c29d-3602-4898-b1a1-b77fa23b8e50|David Bowie]]"
 
* Album "The Wall" by "[[Artist:83d91898-7763-47d7-b03b-b92132375c47|Pink Floyd]]"
 
* Album "The Wall" by "[[Artist:83d91898-7763-47d7-b03b-b92132375c47|Pink Floyd]]"
  +
  +
The table <code>artist_release_group</code> is [https://github.com/metabrainz/musicbrainz-server/blob/88e3d5ce55179aa23e9deccf171e4e1a77efe46e/INSTALL.md#build-materialized-tables materialized (m)].
  +
|}
  +
  +
=== [[Series]] ===
  +
  +
{|
  +
|[[Image:series_entity_details.svg|60px|class=zoomable|Tables for the “series” core entity type properties]]
  +
|A series is a sequence of separate release groups, releases, recordings, works or events with a common theme. The theme is usually prominent in the branding of the entities in the series and the individual entities will often have been given a number indicating the position in the series.
  +
|}
   
 
=== [[URL]] ===
 
=== [[URL]] ===
  +
This entity represents a URL pointing to a resource external to MusicBrainz, i.e. an official homepage, a site where music can be acquired, an entry in another database, etc.
 
  +
{|
  +
|[[Image:url_entity_details.svg|60px|class=zoomable|Tables for the “URL” core entity type properties]]
  +
|This entity represents a URL pointing to a resource external to MusicBrainz, i.e. an official homepage, a site where music can be acquired, an entry in another database, etc.
  +
|}
   
 
=== [[Work]] ===
 
=== [[Work]] ===
   
  +
{|
One layer above ''recording''s ("song", "composition", etc.). While a ''recording'' represents audio data, a ''work'' represents the composition behind the ''recording''. Advanced Relationships are used to link ''recording''s and ''work''s.
 
  +
|[[Image:work_entity_details.svg|60px|class=zoomable|Tables for the “work” core entity type properties]]
  +
|One layer above ''recording''s ("song", "composition", etc.). While a ''recording'' represents audio data, a ''work'' represents the composition behind the ''recording''. Relationships are used to link ''recording''s and ''work''s.
   
 
* Song "''Into the Blue''" by "''Moby''" -- all the recordings listed above will be linked to this object
 
* Song "''Into the Blue''" by "''Moby''" -- all the recordings listed above will be linked to this object
  +
|}
   
 
== Secondary entities ==
 
== Secondary entities ==
Line 74: Line 160:
 
This entity is not visible to users on its own, only in the context of a release. It has an MBID, and contains a link to a recording, a title, artist credit and position on its associated medium.
 
This entity is not visible to users on its own, only in the context of a release. It has an MBID, and contains a link to a recording, a title, artist credit and position on its associated medium.
   
  +
== Entity complementary data ==
== Schema ==
 
   
  +
=== Aliases ===
The SQL scripts that create the schema can be found in our [https://github.com/metabrainz/musicbrainz-server/tree/master/admin/sql source code repository].
 
  +
All the primary entities except URL (i.e. area, artist, event, genre, instrument, label, place, recording, release, release group, series and work) have <code>*_alias</code> tables, all of which have the same structure. They contain alternate names for instances of those entities.
   
  +
=== Annotations ===
This schema diagram shows the core database tables and relationships between them. Core entities are blue, mostly-static lists are yellow, and external identifiers are red. There is also a [http://wiki.musicbrainz.org/-/images/5/52/ngs.png higher resolution schema diagram].
 
  +
All the primary entities except URL (i.e. area, artist, event, genre, instrument, label, place, recording, release, release group, series and work) have a corresponding <code>*_annotation</code> table that links entities of that type to entries in the main <code>annotation</code> table which contains the actual text of the annotation, along with the changelog and the identity of the editor who created it.
   
  +
=== Edits ===
[[Image:ngs.png|700px]]
 
  +
Edits to the database are stored in <code>edit</code> table, with the edit notes/comments stored in <code>edit_note</code>. The 13 other <code>edit_*</code> tables are used to link edits to the entity they modify; there is one such table for each primary entity type. They all have the same structure: just two foreign keys, one to the entity table and the other to <code>edit</code> table.
   
=== *_alias tables ===
+
=== IPIs and ISNIs ===
The primary entities ''area'', ''artist'', ''label'', ''place'' and ''work'' have ''_alias'' tables, all of which have the same structure. They contain the alternate names for instances of those entities.
+
The primary entities ''artist'' and ''label'' can have multiple [[IPI]] and [[ISNI]] codes attached to them. These tables <code>*_ipi</code> and <code>*_isni</code> are used to store those mappings. They contain a foreign key pointing to the entity, the actual value of the code, and <code>edits_pending</code> and ''created'' fields with the usual meaning.
   
  +
=== MBID redirects ===
=== annotation table, & the *_annotation tables ===
 
  +
There are 15 <code>*_gid_redirect</code> tables, one for each of the 13 primary entity types, plus ones for ''track'' and ''artist_credit''. They are used to redirect one MBID to another when entities are merged.
All the primary entities except ''url'' (i.e. ''area'', ''artist'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', & ''work'') has a corresponding ''_annotation'' table that links entities of that type to entries in the main ''annotation'' table which contains the actual text of the annotation, along with the changelog and the identity of the editor who created it.
 
   
=== edit table & the edit_* tables ===
+
=== Ratings ===
  +
There is one <code>*_rating_raw</code> table for each of the primary entities that can be rated: artist, event, label, place, recording, release group and work. The <code>*_rating_raw</code> tables all have the same structure, consisting of the actual rating, expressed as an integer between 0 and 100, and two foreign keys, one linked to the associated entity, and the other linked to the <code>editor</code> table, to specifying who made the rating. For privacy reasons, the <code>*_rating_raw</code> tables aren't included in the database dumps.
Edits to the database are stored in the '''edit''' table, with the edit notes/comments stored in '''edit_note'''. The other '''edit_''' tables are used to link edits to the entity they modify; there is one table for each of the 9 primary entities (''area'', ''artist'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', ''url'', & ''work''). They all have the same structure: just two foreign keys, one to the entity table and the other to '''edit''' table.
 
   
  +
The aggregate rating of each entity is stored in the corresponding <code>_meta</code> table, in the '''rating''' field. There is also a '''rating_count''' field that specifies how many ratings have been entered for the entity.
=== *_ipi & *_isni tables ===
 
The primary entities ''artist'' and ''label'' can have multiple [[IPI]] and [[ISNI]] codes attached to them. The these tables are used to store those mappings. They contain a foreign key pointing to the entity, the actual value of the code, and ''edits_pending'' and ''created'' fields with the usual meaning.
 
   
  +
The <code>release_group_meta</code> table also contains other fields, and the ''release'' entity, even though it cannot be rated, still has a <code>release_meta</code> table, which is used to store other data.
=== tag table, & the *_tag and _tag_raw tables ===
 
All but the ''area'' and ''url'' primary entities (i.e. ''artist'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', & ''work'') have ''_tag'' and ''_tag_raw'' tables, with the same structure. Both the ''_tag'' and ''_tag_raw'' tables contain two foreign keys, linked to the associated entity and to the ''tag'' table. The ''_raw_tag'' tables contain a foreign key, '''editor''', which specifies who added the tag, while the ''_tag'' tables instead contain a '''count''' of how many times a tag is applied to a particular entity, and a '''last_updated''' timestamp.
 
   
  +
=== Tags ===
The ''tag'' table contains the actual names of the tags, and a ''ref_count'' (which is not actually updated anymore, and should be removed).
 
  +
All primary entities but the genre and URL (i.e. area, artist, event, instrument, label, place, recording, release, release group, series and work) have <code>*_tag</code> and <code>*_tag_raw</code> tables, with the same structure. These tables contain two foreign keys, linked to the associated entity and to the <code>tag</code> table. The <code>*_tag_raw</code> tables contain a foreign key, '''editor''', which specifies who added the tag, while the <code>*_tag</code> tables instead contain a '''count''' of how many times a tag is applied to a particular entity, and a '''last_updated''' timestamp. For privacy reasons, the <code>*_tag_raw</code> tables aren't included in the database dumps.
   
  +
The <code>tag</code> table contains the actual names of the tags, and a ''ref_count'' indicating how often the tag has been used.
=== *_gid_redirect tables ===
 
There are 10 ''_gid_redirect'' tables, one for each of the 9 primary entities (''area'', ''artist'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', ''url'' & ''work'') plus one for ''track''s. They are used to redirect one MBID to another when entities are merged.
 
   
=== *_rating_raw & *_meta tables ===
+
=== Types ===
  +
<code>*_type</code> tables are simply mappings between strings and ID numbers, representing various sets of types. Areas, artists, events, instruments, labels, places, series and works are the only primary entities with each have a <code>*_type</code> and an <code>*_alias_type</code> table. Release groups have <code>release_group_primary_type</code> and <code>release_group_secondary_type</code> tables. Genres, recordings and releases only have <code>*_alias_type</code> tables, and URLs do not have <code>*_type</code> tables at all.
There are 5 ''_rating_raw'' tables, one for each of the primary entities that can be rated: ''artist'', ''label'', ''recording'', ''release_group'', & ''work''. The ''_rating_raw'' tables all have the same structure, consisting of the actual rating, expressed as an integer between 0 and 100, and two foreign keys, one linked to the associated entity, and the other linked to the ''editor'' table, to specifying who made the rating.
 
   
  +
== Relationship table structure ==
The aggregate rating of each entity is stored in the corresponding ''_meta'' table, in the '''rating''' field. There is also a '''rating_count''' field that specifies how many ratings have been entered for the entity.
 
   
  +
The first below diagram shows minimal foreign keys needed to define a relationship between two entities. [https://musicbrainz.org/relationships/artist-work Artist-Work] is just an example of a relationship type. The main table is highlighted. All other tables are shortened to focus on foreign key constraints.
The ''release_group_meta'' table also contains other fields, and the ''release'' entity, even though it cannot be rated, still has a ''release_meta'' table, which is used to store other data.
 
  +
The second diagram complementarily shows tables for the detailed representation of a link. The main tables are highlighted.
   
  +
[[Image:relationship_overview.svg|x300px|class=zoomable|Tables for relationship connections]]
=== *_type tables ===
 
  +
[[Image:relationship_details.svg|x300px|class=zoomable|Tables for relationship properties]]
There are 12 ''_type'' tables, which are simply mappings between strings and ID numbers, representing various sets of types; ''area''s, ''artist''s, ''label''s, ''place''s, and ''work''s each have a ''_type'' and an ''_alias_type'' table. ''release_group''s have ''release_group_primary_type'' & ''release_group_secondary_type'' tables. ''Recording''s, ''release''s and ''url''s do not have ''_type'' tables.
 
   
  +
=== l_* tables ===
=== Advanced Relationship table structure ===
 
  +
There are tables for every possible combination of the primary entities (''area'', ''artist'', ''event'', ''instrument'', ''genre'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', ''series'', ''url'' and ''work'') all prefixed with ''l_'' and all with the same format. Two of them are shown in the diagram. They contain a field, '''edits_pending''' that is a count of pending changes to the AR, a '''last_updated''' field, and three foreign keys: '''link''' that points back to the associated entry in the ''link'' table, and '''entity0''' and '''entity1''' that point to the associated entry in the corresponding entity table (i.e. '''artist''', '''recording''', '''url''', etc.).
[[Image:ngs-ars.png|right|500px]]
 
==== l_* tables ====
 
There are tables for every possible combination of the 9 primary entities (''area'', ''artist'', ''label'', ''place'', ''recording'', ''release'', ''release_group'', ''url'' & ''work'') all prefixed with ''l_'' and all with the same format. Two of them are shown in the diagram. They contain a field, '''edits_pending''' that is a count of pending changes to the AR, a '''last_updated''' field, and three foreign keys: '''link''' that points back to the associated entry in the ''link'' table, and '''entity0''' and '''entity1''' that point to the associated entry in the corresponding entity table (i.e. '''artist''', '''recording''', '''url''', etc.).
 
   
 
There are two tables that help to avoid unnecessary duplication:
 
There are two tables that help to avoid unnecessary duplication:
   
====link & link_attribute tables====
+
=== link & link_attribute tables ===
   
 
The ''link'' table contains the begin and end date info, and the '''link_type''' foreign key field that specifies what kind of AR it is. It also has a count of how many other attributes that particular link has in the '''attribute_count''' field, and a '''created''' field that specifies when it was created.
 
The ''link'' table contains the begin and end date info, and the '''link_type''' foreign key field that specifies what kind of AR it is. It also has a count of how many other attributes that particular link has in the '''attribute_count''' field, and a '''created''' field that specifies when it was created.
Line 125: Line 211:
 
Each AR attribute either applies to a particular AR or it doesn't. The ''link_attribute'' table stores this information, having a record for each '''attribute_type''' (a foreign key field for ''link_attribute_type'') of each '''link'''.
 
Each AR attribute either applies to a particular AR or it doesn't. The ''link_attribute'' table stores this information, having a record for each '''attribute_type''' (a foreign key field for ''link_attribute_type'') of each '''link'''.
   
====link_type, link_attribute_type, & link_type_attribute_type tables====
+
=== link_type, link_attribute_type, & link_type_attribute_type tables ===
   
 
The AR types and attributes are defined in these three tables, shown at the top of the diagram. They can only be modified by the AR editors.
 
The AR types and attributes are defined in these three tables, shown at the top of the diagram. They can only be modified by the AR editors.
Line 135: Line 221:
 
The ''link_type_attribute_type'' table specifies what attributes can be applied to particular types of ARs; it has the necessary foreign key fields ('''link_type''' and '''attribute_type''') and it also specifies how many instances of the attribute (or one of its children) can be added to the particular AR type in the '''min''' and '''max''' fields. Currently, most of them allow the attributes to merely be present or absent, while a few allow any number of copies of the attribute, or none. The "creative commons licensed download" attribute has to be included exactly once, while the ''instrument'' attribute (of the ''instrument'' type AR), requires at least one instance.
 
The ''link_type_attribute_type'' table specifies what attributes can be applied to particular types of ARs; it has the necessary foreign key fields ('''link_type''' and '''attribute_type''') and it also specifies how many instances of the attribute (or one of its children) can be added to the particular AR type in the '''min''' and '''max''' fields. Currently, most of them allow the attributes to merely be present or absent, while a few allow any number of copies of the attribute, or none. The "creative commons licensed download" attribute has to be included exactly once, while the ''instrument'' attribute (of the ''instrument'' type AR), requires at least one instance.
   
===Cover Art Archive table structure===
+
== Cover Art Archive table structure ==
  +
{|
[[Image:cover art.png|right|400px]]
 
  +
|[[Image:cover_art_details.svg|x300px|class=zoomable|Tables for cover art]]
 
The [[Cover Art Archive]] table structure is fairly simple. The ''cover_art'' table stores the actual cover art and associations to edits. ''art_type'' stores the acceptable cover art types, and ''cover_art_type'' links the two together. ''release_group_cover_art'' links a release group to the release whose cover art should represent the release group.
+
|The [[Cover Art Archive]] table structure is fairly simple. The ''cover_art'' table stores the actual cover art and associations to edits. ''art_type'' stores the acceptable cover art types, and ''cover_art_type'' links the two together. ''release_group_cover_art'' links a release group to the release whose cover art should represent the release group.
   
 
This image also shows the links to tables in the main diagram (as well as the ''edit'' table), but not their full schemas; please see other diagrams or the real schema specification for details.
 
This image also shows the links to tables in the main diagram (as well as the ''edit'' table), but not their full schemas; please see other diagrams or the real schema specification for details.
   
 
There is one view not shown, which is the ''index_listing'' view -- this makes for a slightly nicer interface than joining the tables manually, by providing an array of cover art types and easy pointers for is_front and is_back. Otherwise it largely resembles the cover_art table.
 
There is one view not shown, which is the ''index_listing'' view -- this makes for a slightly nicer interface than joining the tables manually, by providing an array of cover art types and easy pointers for is_front and is_back. Otherwise it largely resembles the cover_art table.
  +
|}
   
===CDStubs table structure===
+
== CDStubs table structure ==
  +
{|
[[Image:cdstubs.png|right|300px]]
 
  +
|[[Image:cdstub_details.svg|x300px|class=zoomable|Tables for CD stub]]
 
The [[CDStub]]s table structure consists of only 3 tables: cdtoc_raw, release_raw and track_raw.
+
|The [[CDStub]]s table structure consists of only 3 tables: cdtoc_raw, release_raw and track_raw.
 
The ''release_raw'' table contains basic metadata about the stub release, such as the artist and title.
 
The ''release_raw'' table contains basic metadata about the stub release, such as the artist and title.
   
The individual tracks of each stub release are in the ''track_raw'' table, which holds the track title and its track number (in the ''sequence'' column). A track's artist can be different than the artist of the release, so there's an ''artist'' column which can optionally contain the name of the tracks artist.
+
The individual tracks of each stub release are in the ''track_raw'' table, which holds the track title and its track number (in the ''sequence'' column). A track's artist can be different than the artist of the release, so there's an ''artist'' column which can optionally contain the name of the track's artist.
   
 
Each CDStub needs an associated [[Disc ID]], which is stored in the ''cdtoc_raw'' table, together with the track count of the disc, as well as the leadout and track offset ([[Disc_ID_Calculation#Audio_CD|Disc ID Calculation]] contains more information about the latter two).
 
Each CDStub needs an associated [[Disc ID]], which is stored in the ''cdtoc_raw'' table, together with the track count of the disc, as well as the leadout and track offset ([[Disc_ID_Calculation#Audio_CD|Disc ID Calculation]] contains more information about the latter two).
  +
|}
  +
  +
== Frequent use cases ==
  +
  +
=== Tagging audio files ===
  +
{|
  +
|[[Image:for_tagging_audio_files.svg|x100px|class=zoomable|Tables for tagging audio files]]
  +
|When roughly tagging audio files with basic metadata, you will mainly be looking for:
  +
* Primary entity types: artist, recording, release, release group
  +
* Secondary entity types: artist credit, medium, track
  +
|}
  +
  +
=== Finding song authors ===
  +
{|
  +
|[[Image:for_finding_song_authors.svg|x100px|class=zoomable|Tables for finding song authors]]
  +
|When retrieving the authors (composer, lyricist…) of a recorded song, you will mainly be looking for:
  +
* Primary entity types: artist, recording, work
  +
* Secondary entity types: artist credit
  +
* Relationship types: artist-work, recording-work
  +
|}
   
 
== Undocumented tables ==
 
== Undocumented tables ==
Line 158: Line 265:
   
 
<pre>
 
<pre>
CREATE TABLE application
+
CREATE TABLE artist_credit_name
 
CREATE TABLE artist_deletion
 
 
CREATE TABLE artist_credit_name (
 
 
CREATE TABLE autoeditor_election
 
CREATE TABLE autoeditor_election_vote
 
   
 
CREATE TABLE cdtoc
 
CREATE TABLE cdtoc
CREATE TABLE cdtoc_raw
 
   
CREATE TABLE editor
+
CREATE TABLE gender
CREATE TABLE editor_language (
 
CREATE TABLE editor_preference
 
CREATE TABLE editor_subscribe_artist
 
CREATE TABLE editor_subscribe_artist_deleted
 
CREATE TABLE editor_subscribe_collection
 
CREATE TABLE editor_subscribe_label
 
CREATE TABLE editor_subscribe_label_deleted
 
CREATE TABLE editor_subscribe_editor
 
   
CREATE TABLE gender (
+
CREATE TABLE iso_3166_1
  +
CREATE TABLE iso_3166_2
 
CREATE TABLE iso_3166_1 (
+
CREATE TABLE iso_3166_3
CREATE TABLE iso_3166_2 (
 
CREATE TABLE iso_3166_3 (
 
   
 
CREATE TABLE isrc
 
CREATE TABLE isrc
CREATE TABLE iswc (
+
CREATE TABLE iswc
 
CREATE TABLE label_deletion
 
   
 
CREATE TABLE language
 
CREATE TABLE language
   
CREATE TABLE link_creditable_attribute_type (
+
CREATE TABLE link_creditable_attribute_type
CREATE TABLE link_attribute_credit (
+
CREATE TABLE link_attribute_credit
 
CREATE TABLE editor_collection
 
CREATE TABLE editor_collection_release
 
CREATE TABLE editor_oauth_token
 
CREATE TABLE editor_watch_preferences
 
CREATE TABLE editor_watch_artist
 
CREATE TABLE editor_watch_release_group_type
 
CREATE TABLE editor_watch_release_status
 
   
 
CREATE TABLE medium_cdtoc
 
CREATE TABLE medium_cdtoc
 
CREATE TABLE medium_format
 
CREATE TABLE medium_format
   
CREATE TABLE replication_control
+
CREATE TABLE medium_index
   
CREATE TABLE release_country (
+
CREATE TABLE release_country
CREATE TABLE release_unknown_country (
+
CREATE TABLE release_unknown_country
CREATE TABLE release_raw
 
 
CREATE TABLE release_coverart
 
CREATE TABLE release_coverart
CREATE TABLE release_label (
+
CREATE TABLE release_label
 
CREATE TABLE release_packaging
 
CREATE TABLE release_packaging
 
CREATE TABLE release_status
 
CREATE TABLE release_status
   
CREATE TABLE release_group_secondary_type_join (
+
CREATE TABLE release_group_secondary_type_join
  +
  +
CREATE TABLE replication_control
   
 
CREATE TABLE script
 
CREATE TABLE script
Line 224: Line 304:
 
CREATE TABLE tag_relation
 
CREATE TABLE tag_relation
   
CREATE TABLE track_raw
+
CREATE TABLE work_attribute_type
  +
CREATE TABLE work_attribute_type_allowed_value
  +
CREATE TABLE work_attribute
  +
</pre>
   
  +
The following tables are not publicly available for download thus are not relevant in this page.
CREATE TABLE medium_index
 
   
  +
<pre>
CREATE TABLE vote
 
  +
CREATE TABLE application
   
CREATE TABLE work_attribute_type (
+
CREATE TABLE autoeditor_election
CREATE TABLE work_attribute_type_allowed_value (
+
CREATE TABLE autoeditor_election_vote
  +
CREATE TABLE work_attribute (
 
  +
CREATE TABLE edit_note_change
  +
  +
CREATE TABLE editor
  +
CREATE TABLE editor_collection
  +
CREATE TABLE editor_collection_release
  +
CREATE TABLE editor_language
  +
CREATE TABLE editor_oauth_token
  +
CREATE TABLE editor_preference
  +
CREATE TABLE editor_subscribe_artist
  +
CREATE TABLE editor_subscribe_artist_deleted
  +
CREATE TABLE editor_subscribe_collection
  +
CREATE TABLE editor_subscribe_label
  +
CREATE TABLE editor_subscribe_label_deleted
  +
CREATE TABLE editor_subscribe_editor
  +
CREATE TABLE editor_subscribe_series
  +
CREATE TABLE editor_subscribe_series_deleted
  +
  +
CREATE TABLE unreferenced_row_log
  +
  +
CREATE TABLE vote
 
</pre>
 
</pre>
   

Latest revision as of 18:44, 21 July 2023

Status: This page is a work in progress.

This page is intended for developers querying the MusicBrainz database directly through PostgreSQL.

Overview

The SQL scripts that create the schema can be found in our source code repository.

The database is structured around primary entities which can be edited, searched for, referred to by MBID, and linked to each other through relationships or foreign key constraints in some case. Secondary entities cannot be linked through relationships, instead they are directly linked through foreign key constraints only. Each primary entity type has a main table sometimes accompanied by complementary tables of which a few are unique but most are common.

  • The first diagram below shows the main tables for primary entity types and unique tables connecting these tables, that is, beyond relationships. The main tables are highlighted. All tables are shortened to focus on foreign key constraints.
  • The second diagram additionally shows the main tables for primary entity types and tables unique to some of these, that is, beyond aliases, annotations, edits, redirects, relationships, ratings, tags.
  • The third diagram shows all the tables but is too dense to just distinguish foreign key constraints for example, thus the many following sections.

Overview of tables connecting core entities Tables specifically related to core entities All tables (exhaustive but indigestible)

Primary entities

Each primary entity has a main table with the same name as the entity, containing its basic data.

Area

Tables for the “area” core entity type properties A country, region, city or the like.

Areas that can be used for filling in the Release country field of releases are listed, by ID, in the country_area table.

The table area_containment is materialized (m).

Artist

Tables for the “artist” core entity type properties An artist is generally a musician, a group of musicians, or another music professional (composer, engineer, illustrator, producer, etc.)

The tables artist_release and artist_release_group are materialized (m).

Event

Tables for the “event” core entity type properties An event refers to an organised event which people can attend, and is relevant to MusicBrainz. Generally this means live performances, like concerts and festivals.

Genre

Tables for the “genre” core entity type properties A genre is a descriptor for the style and conventions followed by a piece of music.

Instrument

Tables for the “instrument” core entity type properties Instruments are devices created or adapted to make musical sounds. We also list common instrument groupings (such as string quartet) as instruments.

Label

Tables for the “label” core entity type properties Labels represent mostly (but not only) imprints.

Place

Tables for the “place” core entity type properties A venue, studio or other place where music is performed, recorded, engineered, etc.

Recording

Tables for the “recording” core entity type properties Represents a unique mix or edit. Has title, artist credit, duration, list of ISRCs. Examples (all are different Recordings):
  • Album version of the track "Into the Blue" by "Moby"
  • Remix "Into the Blue (Buzz Boys Main Room Mayhem mix)" by "Moby"
  • Remix "Into the Blue (Underground mix)" by "Moby"

The table recording_first_release_date is materialized (m).

Release

Tables for the “release” core entity type properties Real-world release object you can buy in your music store. It has release date and country, list of catalog number and label pairs, packaging type and release status. Examples:
  • 1984 US release of "The Wall" by "Pink Floyd", release on label "Columbia Records" with catalog number "C2K 36183" and UPC "074643618328", it's an official release and comes with two CDs in jewel case.

The tables artist_release and release_first_release_date are materialized (m).

Release group

Tables for the “release group” core entity type properties Represents an abstract "album" (or "single", or "EP") entity. Technically it's a group of releases, with a specified type. Examples:

The table artist_release_group is materialized (m).

Series

Tables for the “series” core entity type properties A series is a sequence of separate release groups, releases, recordings, works or events with a common theme. The theme is usually prominent in the branding of the entities in the series and the individual entities will often have been given a number indicating the position in the series.

URL

Tables for the “URL” core entity type properties This entity represents a URL pointing to a resource external to MusicBrainz, i.e. an official homepage, a site where music can be acquired, an entry in another database, etc.

Work

Tables for the “work” core entity type properties One layer above recordings ("song", "composition", etc.). While a recording represents audio data, a work represents the composition behind the recording. Relationships are used to link recordings and works.
  • Song "Into the Blue" by "Moby" -- all the recordings listed above will be linked to this object

Secondary entities

Each secondary entity has a main table with the same name as the entity, containing its basic data.

Artist credit

List of artists, variations of artist names and pieces of text to join the artist names. Examples:

Medium

This entity represents a piece of media, included in a release. It contains information about the format of the media, its position in the release, an optional title, and most importantly, a list of tracks. CD TOCs are attached to mediums, not releases or release_groups.

Examples:

  • CD1 of the 1984 US release of "The Wall" by "Pink Floyd"
  • CD2 of the 2005 UK release of "Aerial" by "Kate Bush", named "A Sky of Honey"

Track

This entity is not visible to users on its own, only in the context of a release. It has an MBID, and contains a link to a recording, a title, artist credit and position on its associated medium.

Entity complementary data

Aliases

All the primary entities except URL (i.e. area, artist, event, genre, instrument, label, place, recording, release, release group, series and work) have *_alias tables, all of which have the same structure. They contain alternate names for instances of those entities.

Annotations

All the primary entities except URL (i.e. area, artist, event, genre, instrument, label, place, recording, release, release group, series and work) have a corresponding *_annotation table that links entities of that type to entries in the main annotation table which contains the actual text of the annotation, along with the changelog and the identity of the editor who created it.

Edits

Edits to the database are stored in edit table, with the edit notes/comments stored in edit_note. The 13 other edit_* tables are used to link edits to the entity they modify; there is one such table for each primary entity type. They all have the same structure: just two foreign keys, one to the entity table and the other to edit table.

IPIs and ISNIs

The primary entities artist and label can have multiple IPI and ISNI codes attached to them. These tables *_ipi and *_isni are used to store those mappings. They contain a foreign key pointing to the entity, the actual value of the code, and edits_pending and created fields with the usual meaning.

MBID redirects

There are 15 *_gid_redirect tables, one for each of the 13 primary entity types, plus ones for track and artist_credit. They are used to redirect one MBID to another when entities are merged.

Ratings

There is one *_rating_raw table for each of the primary entities that can be rated: artist, event, label, place, recording, release group and work. The *_rating_raw tables all have the same structure, consisting of the actual rating, expressed as an integer between 0 and 100, and two foreign keys, one linked to the associated entity, and the other linked to the editor table, to specifying who made the rating. For privacy reasons, the *_rating_raw tables aren't included in the database dumps.

The aggregate rating of each entity is stored in the corresponding _meta table, in the rating field. There is also a rating_count field that specifies how many ratings have been entered for the entity.

The release_group_meta table also contains other fields, and the release entity, even though it cannot be rated, still has a release_meta table, which is used to store other data.

Tags

All primary entities but the genre and URL (i.e. area, artist, event, instrument, label, place, recording, release, release group, series and work) have *_tag and *_tag_raw tables, with the same structure. These tables contain two foreign keys, linked to the associated entity and to the tag table. The *_tag_raw tables contain a foreign key, editor, which specifies who added the tag, while the *_tag tables instead contain a count of how many times a tag is applied to a particular entity, and a last_updated timestamp. For privacy reasons, the *_tag_raw tables aren't included in the database dumps.

The tag table contains the actual names of the tags, and a ref_count indicating how often the tag has been used.

Types

*_type tables are simply mappings between strings and ID numbers, representing various sets of types. Areas, artists, events, instruments, labels, places, series and works are the only primary entities with each have a *_type and an *_alias_type table. Release groups have release_group_primary_type and release_group_secondary_type tables. Genres, recordings and releases only have *_alias_type tables, and URLs do not have *_type tables at all.

Relationship table structure

The first below diagram shows minimal foreign keys needed to define a relationship between two entities. Artist-Work is just an example of a relationship type. The main table is highlighted. All other tables are shortened to focus on foreign key constraints. The second diagram complementarily shows tables for the detailed representation of a link. The main tables are highlighted.

Tables for relationship connections Tables for relationship properties

l_* tables

There are tables for every possible combination of the primary entities (area, artist, event, instrument, genre, label, place, recording, release, release_group, series, url and work) all prefixed with l_ and all with the same format. Two of them are shown in the diagram. They contain a field, edits_pending that is a count of pending changes to the AR, a last_updated field, and three foreign keys: link that points back to the associated entry in the link table, and entity0 and entity1 that point to the associated entry in the corresponding entity table (i.e. artist, recording, url, etc.).

There are two tables that help to avoid unnecessary duplication:

link & link_attribute tables

The link table contains the begin and end date info, and the link_type foreign key field that specifies what kind of AR it is. It also has a count of how many other attributes that particular link has in the attribute_count field, and a created field that specifies when it was created.

Each AR attribute either applies to a particular AR or it doesn't. The link_attribute table stores this information, having a record for each attribute_type (a foreign key field for link_attribute_type) of each link.

link_type, link_attribute_type, & link_type_attribute_type tables

The AR types and attributes are defined in these three tables, shown at the top of the diagram. They can only be modified by the AR editors.

The link_type table defines the types of ARs available. AR types are arranged in a number of trees, for ease of finding. This tree structure is expressed with the parent and child_order fields; parent is the id of the parent AR type, or null if it's at the root, and child_order orders the children of a given parent AR type. Each AR type has a unique uuid, stored in the gid field, for use in permalinks and external applications. The link between a particular AR type and the corresponding l_ table is formed by the entity_type0 and entity_type1 fields.

The attributes are themselves defined in the link_attribute_type table. Like AR types, attributes form a number of trees (the vast majority of them are individual musical instruments). Besides the parent and child_order fields shared with the link_type table, the link_attribute_table also has a root field, showing the root of the tree that the attribute is part of. Attributes also have names and descriptions which appear in various places where they are displayed, as well as gids and a last_updated timestamp.

The link_type_attribute_type table specifies what attributes can be applied to particular types of ARs; it has the necessary foreign key fields (link_type and attribute_type) and it also specifies how many instances of the attribute (or one of its children) can be added to the particular AR type in the min and max fields. Currently, most of them allow the attributes to merely be present or absent, while a few allow any number of copies of the attribute, or none. The "creative commons licensed download" attribute has to be included exactly once, while the instrument attribute (of the instrument type AR), requires at least one instance.

Cover Art Archive table structure

Tables for cover art The Cover Art Archive table structure is fairly simple. The cover_art table stores the actual cover art and associations to edits. art_type stores the acceptable cover art types, and cover_art_type links the two together. release_group_cover_art links a release group to the release whose cover art should represent the release group.

This image also shows the links to tables in the main diagram (as well as the edit table), but not their full schemas; please see other diagrams or the real schema specification for details.

There is one view not shown, which is the index_listing view -- this makes for a slightly nicer interface than joining the tables manually, by providing an array of cover art types and easy pointers for is_front and is_back. Otherwise it largely resembles the cover_art table.

CDStubs table structure

Tables for CD stub The CDStubs table structure consists of only 3 tables: cdtoc_raw, release_raw and track_raw.

The release_raw table contains basic metadata about the stub release, such as the artist and title.

The individual tracks of each stub release are in the track_raw table, which holds the track title and its track number (in the sequence column). A track's artist can be different than the artist of the release, so there's an artist column which can optionally contain the name of the track's artist.

Each CDStub needs an associated Disc ID, which is stored in the cdtoc_raw table, together with the track count of the disc, as well as the leadout and track offset (Disc ID Calculation contains more information about the latter two).

Frequent use cases

Tagging audio files

Tables for tagging audio files When roughly tagging audio files with basic metadata, you will mainly be looking for:
  • Primary entity types: artist, recording, release, release group
  • Secondary entity types: artist credit, medium, track

Finding song authors

Tables for finding song authors When retrieving the authors (composer, lyricist…) of a recorded song, you will mainly be looking for:
  • Primary entity types: artist, recording, work
  • Secondary entity types: artist credit
  • Relationship types: artist-work, recording-work

Undocumented tables

The following tables have not yet been documented in this page. Help is gratefully appreciated!

CREATE TABLE artist_credit_name

CREATE TABLE cdtoc

CREATE TABLE gender

CREATE TABLE iso_3166_1
CREATE TABLE iso_3166_2
CREATE TABLE iso_3166_3

CREATE TABLE isrc
CREATE TABLE iswc

CREATE TABLE language

CREATE TABLE link_creditable_attribute_type
CREATE TABLE link_attribute_credit

CREATE TABLE medium_cdtoc
CREATE TABLE medium_format

CREATE TABLE medium_index

CREATE TABLE release_country
CREATE TABLE release_unknown_country
CREATE TABLE release_coverart
CREATE TABLE release_label
CREATE TABLE release_packaging
CREATE TABLE release_status

CREATE TABLE release_group_secondary_type_join

CREATE TABLE replication_control

CREATE TABLE script
CREATE TABLE script_language

CREATE TABLE tag_relation

CREATE TABLE work_attribute_type
CREATE TABLE work_attribute_type_allowed_value
CREATE TABLE work_attribute

The following tables are not publicly available for download thus are not relevant in this page.

CREATE TABLE application

CREATE TABLE autoeditor_election
CREATE TABLE autoeditor_election_vote

CREATE TABLE edit_note_change

CREATE TABLE editor
CREATE TABLE editor_collection
CREATE TABLE editor_collection_release
CREATE TABLE editor_language
CREATE TABLE editor_oauth_token
CREATE TABLE editor_preference
CREATE TABLE editor_subscribe_artist
CREATE TABLE editor_subscribe_artist_deleted
CREATE TABLE editor_subscribe_collection
CREATE TABLE editor_subscribe_label
CREATE TABLE editor_subscribe_label_deleted
CREATE TABLE editor_subscribe_editor
CREATE TABLE editor_subscribe_series
CREATE TABLE editor_subscribe_series_deleted

CREATE TABLE unreferenced_row_log

CREATE TABLE vote