Simple hierarchical Example[edit]

9 single entries for 2022 sorted by social media followers[edit]

9 single entries for 2022 sorted by social media followers[edit]

SPARQL query[edit]

Hide/Show SPARQLquerySingles2022

# WF 2022-06-08
# Singles published in 2022
SELECT
?single ?language ?collectionId ?collection ?collectionType ?performerId ?performer ?followerCount ?youtubeVideoId  (year(?publicationDate) as ?publicationYear)
WHERE
{
  # instanceof
  ?singleId wdt:P31 wd:Q134556. # Must be of a single
  ?singleId rdfs:label ?single.
  FILTER(LANG(?single)='en')

  #https://www.wikidata.org/wiki/Property:P361 - part of / ist Teil von
  ?singleId wdt:P361 ?collectionId.
  ?collectionId rdfs:label ?collection.
  FILTER(LANG(?collection)='en')

  # https://www.wikidata.org/wiki/Property:P407 language/ Sprache des Werkes
  ?singleId wdt:P407 ?languageId.
  ?languageId rdfs:label ?language.
  FILTER(LANG(?language)='en')

  # https://www.wikidata.org/wiki/Property:P1651 YouTube-Video-Kennung
  ?singleId wdt:P1651 ?youtubeVideoId.

  ?collectionId wdt:P31 ?collectionTypeId.
  ?collectionTypeId rdfs:label ?collectionType.
  FILTER(LANG(?collectionType)='en').

  # https://www.wikidata.org/wiki/Property:P175 Performer /Interpret
  ?singleId wdt:P175 ?performerId.
  ?performerId rdfs:label ?performer
  FILTER(LANG(?performer)='en')

  # https://www.wikidata.org/wiki/Property:P8687 followers / Follower
  #{
     #SELECT (max(?followerCount) as ?follower)
     #WHERE {
     ?performerId wdt:P8687 ?followerCount.
     #}
     #GROUP BY ?performerId
  #}
  #https://www.wikidata.org/wiki/Property:P577 Publication date / Veröffentlichungsdatum
  ?singleId wdt:P577 ?publicationDate.
  FILTER(year(?publicationDate) = 2022).
} ORDER BY  DESC(?followerCount)
LIMIT 9

try it!

result as of 2022-06-13[edit]

Functional Dependencies[edit]

The functional dependencies in this example are derived from the structure of the SPARQL query

• A≡single
• B≡language
• C≡collectionId
• D≡collection
• E≡collectionType
• F≡performerId
• G≡performer
• H≡followerCount
• I≡youtubeVideoId
• J≡publicationYear

• [math]A\to BCFIJ[/math]
• [math]C\to DE[/math]
• [math]F\to GH[/math]

[math]R={{\{A,B,C,D,E,F,G,H,I,J\}},{A\to BCFIJ,C\to DE,F\to GH}}[/math]

functional dependencies diagram[edit]

The following diagram shows the functional dependencies of this example. The nodes represent the attributes and the edges the dependencies.

Attribute Closures[edit]

The attribute closure is the set of all attributes that are "reachable/determinable" from a given attribute node. In this example the results are quite obvious since the functional dependencies form a tree and the reachable nodes are all nodes of the subtree of a given attribute.

Keys for the example[edit]

There is only a single candidate key: A≡single

It is the only attribute that has the full set of attributes as it's closure. A≡single is also a super key and primary key

With a slightly modified query we'll see that wikidata uses a surrogate key as the primary key for all items. Querying Wikidata's graph data more often than not leads to results that need to be normalized to be useful in the context of relational databases.

Extending the above query with a column for the wikidata Id of the single shows how the surrogate key is related to the single entries.

9 single entries for 2022 sorted by social media followers with wikidata Id link[edit]

query[edit]

Hide/Show SPARQL_Query_With_WikidataId

# WF 2022-06-08
# Singles published in 2022
SELECT
?singleId ?single ?language ?collectionId ?collection ?collectionType ?performerId ?performer ?followerCount ?youtubeVideoId  (year(?publicationDate) as ?publicationYear)
WHERE
{
  # instanceof
  ?singleId wdt:P31 wd:Q134556. # Must be of a single
  ?singleId rdfs:label ?single.
  FILTER(LANG(?single)='en')

  #https://www.wikidata.org/wiki/Property:P361 - part of / ist Teil von
  ?singleId wdt:P361 ?collectionId.
  ?collectionId rdfs:label ?collection.
  FILTER(LANG(?collection)='en')

  # https://www.wikidata.org/wiki/Property:P407 language/ Sprache des Werkes
  ?singleId wdt:P407 ?languageId.
  ?languageId rdfs:label ?language.
  FILTER(LANG(?language)='en')

  # https://www.wikidata.org/wiki/Property:P1651 YouTube-Video-Kennung
  ?singleId wdt:P1651 ?youtubeVideoId.

  ?collectionId wdt:P31 ?collectionTypeId.
  ?collectionTypeId rdfs:label ?collectionType.
  FILTER(LANG(?collectionType)='en').

  # https://www.wikidata.org/wiki/Property:P175 Performer /Interpret
  ?singleId wdt:P175 ?performerId.
  ?performerId rdfs:label ?performer
  FILTER(LANG(?performer)='en')

  # https://www.wikidata.org/wiki/Property:P8687 followers / Follower
  #{
     #SELECT (max(?followerCount) as ?follower)
     #WHERE {
     ?performerId wdt:P8687 ?followerCount.
     #}
     #GROUP BY ?performerId
  #}
  #https://www.wikidata.org/wiki/Property:P577 Publication date / Veröffentlichungsdatum
  ?singleId wdt:P577 ?publicationDate.
  FILTER(year(?publicationDate) = 2022).
} ORDER BY  DESC(?followerCount)
LIMIT 9

try it!

result[edit]

Normal forms[edit]

1st Normal Form - 1NF[edit]

A relation is in first normal form if and only if no attribute domain has relations as elements Wikipedia Article on First normal form

• attribute domain means the type of attribute (think of attribute a cell of a table in spreadsheet and the domain as the type you selected for the cells of a column)
• the type not having releations is sometime also describe as "atomic" meaning having a base type such a string, float, integer and not tuples
• an example for a cell having a relations a elements would be having LAT and LON of a geocoordinate in the same cell
• if you have lists of values or even full tables in a cell than the relation is surely not in 1st normal Form 1NF
• if a single cell in a column violates 1NF the whole columns type is violating 1NF which makes the whole table violate 1NF
• in modern systems where you want to be lenient with user input its often hard to enforce 1NF - database designers might try to "cutting corners" around this by treating the type

as string event if it contains relations data such as a full list of comma separated elements, a json or xml encoded or even uuencoded binary string.

The example above is in first normal form the types of the columns are:

• string: A≡single, B≡language, D≡collection, E≡collectionType, G≡performer, I≡youtubeVideoId
• integer: H≡followerCount, J≡publicationYear
• url: C≡collectionId, F≡performerId

2nd Normal Form[edit]

3rd Normal Form[edit]

Boyce Codd Normal Form[edit]