Ethnomusicology Database

This document presents the structures of the original 4D database of the CNRS Ethnomusicology laboratory (LEM).

4D database structure

Diagram showing the different tables and their relations:

Data tables

The Ethnomusicology database contains tables with obsolete/obscure names. We also need to translate them from french to english. Additionally, some tables are themselves obsolete, and some contain unstandardized informations which we might replace with ISO standards and the like.

old table name	new table name	comment
Support	collection
Phono	item
Ethnie	ethnic_group
Alias_Ethnie	?	These are aliases for the ethnic groups. Not sure what to do with it.
Compt_Ethnie	?	What's this?
Fonction_Usage	?
Mot_Clef	?
Compt_CotesSup	?
Alias_CotesSup	?
Autres_Cotes	?
Etat	?	This is an non-standardized list of countries. We might drop this table. Dublin Core recommends to use TGN for geographic coverage. We migh also use ISO-3166.
Alias_Etat	?	This is a list of state aliases. We might drop it.
Compt_Etat	?	What's this?
Scientifique	scientific_instrument	This is a list of scientific instrument names. Don't know if there's a standard for this
Vernaculaire	vernacular_instrument	A list of instrument names in native languages
Form	?	What's this?
Modale	?	Obsolete?
Rythm	?	Obsolete?
Param	?	4D-specific?
Images	?	4D-specific?

Collection fields

These are fields for the collection table (old table name: Support).

old field name	new field name	Type	Dublin Core mapping	comment
Réf	publisher_reference	String		This field contains the reference assigned by a publishing company, when the collection has been published. If it hasn't been published, it contains a reference assigned by the LEM, which somewhat duplicates the information contained in the field "Cote". In this later case, it has the form BM.XXX.XXX, where X are digits (note: some published references also start by "BM", but have a different form)
Format	physical_format	Enumeration		Format of the original physical support, historical data. Doesn't really appear suitable for Dublin Core's format
Cote	id	String	identifier	This is a unique identifier for the collection. It is currently of the form "BM.YYY.NNN" where YYY is the year and NNN a serial number. It will evolve to CNRSMH_I_YYYY_NNN
Titre	title	String	title
Transcript_trad	native_title	String	title	This is either the translation from the native title, or its phonetic transcription. We might simply consider it as an alternative title, this is Dublin Core compliant, a resource can have several titles.
Nb_de_pieces	physical_items_num	Integer		If the collection is on a magnetic tape this is the number of tapes. If it is on a disk it is the number of tracks. This is historical data. We can keep it, but will also display a more reliable computed number of related items.
Réédition	publishing_status	Enumeration		Wether the collection was fully/partially published, reedited, or not
Original	is_original	Boolean		Whether this is an genuine/original collection, as opposed to copied ones
Copie_TotPartie	is_full_copy	Boolean		If this is not an original collection (See the is_original field), then whether the collection was partially or fully copied. Note: this field is used for only 2 records on more than 4000
Copié_de	copied_from	Foreign key (recursive)	relation	Indicates which collection this collection has been (fully or partially) copied from.
Auteur_compil	creator	String	creator	The person who created this collection, by choosing which items to group together
Auteur_notice	booklet_writer	String	contributor	The person who wrote the disk booklet (only for disk booklets?). Empty if the collection is not published.
Notice	booklet_description	Text		Description of the technical documentation (= disk booklet?): its number of pages, languages, etc...
Collecteur	collector	String	contributor	The person who actually recorded the audio items. If the same as the creator (often), contains "="
Editeur	publisher	Enumeration	publisher	This is the external publishing company. If not published, might contain mixed data: "Original*", "Copie*", etc... This extra data is redundant with the is_original and is_full_copy fields, and should/might be removed.
Année parution	date_published	Integer	date	This field contains -1 if the year is unknown, and -2 if the collection hasn't been published. We might map it to a proper SQL DATE field.
Collect_série	publisher_collection	Enumeration		The name of the publisher-specific "collection"
Num_dans_collec	publisher_serial_id	Integer		Publisher-specific serial id within its "collection"
réf_biblio	ref_biblio	String
mode_acquisition	acquisition_mode	Enumeration		The way the collection got acquired by the LEM: gift, buying, etc...
commentaires	comment	Text		Many things in there...
Redacteur_fiche	record_author	String		The person who authored this collection metadata
Saisie_fiche	record_writer	String		The person who actually wrote this collection metadata. In telemeta this field might be automatically filled with the currently logged in user.

Item fields

These are fields for the item table (old table name: Phono).

old field name	new field name	Type	Dublin Core mapping	comment
Cote_Phono	id	String	identifier	This is a unique identifier for the item. It currently contains the collection id ("BM.YYY.NNN") suffixed with some rather messy serial number followed by track number(s), etc.. It will evolve to CNRSMH_I_YYYY_NNN_MMM_TT_PP, where MMM is the number of the original tape, TT the track number and PP the part number. Remark: if some ids to not contain the PP suffix (this is expected), it should be assumed that PP = 01.

Resources

• An interesting forum thread about exporting a 4D database to Excel/CSV and other 4D versions (french). Contains links to freely downloadable 4D demo versions, that seem usable.