Fetching contributors…
Cannot retrieve contributors at this time
555 lines (482 sloc) 17.8 KB

Read operations


wd summary

wd summary Q1

Working with Wikidata, we often end up with obscure ids. We can always look-up those ids labels on the website but that means loading pages and pages, when a small API call and parsing could return just what we need: a label, a description, and some claims to know what we are dealing with.

wd summary <entities ids>
# Alias:
wd s <entities ids>
wd summary Q27477672
# Label anthropomorphic comic
# Description comic book genre
# instance of (P31):  comic genre (Q20087698)
# subclass of (P279): animal comic (Q10493450)
wd summary Q18120925 Q22117436 Q22117437


  • -p, --properties <properties>: request additional properties (separated by a comma)
  • -l, --lang <lang>: specify the summary's language
  • -v, --verbose: log all claims

wd search

wd search <query>
# Alias: (the s was already used by 'wd summary', so 'f' as 'find')
wd f <query>
wd search Ligo
# Q255371    Laser — Interferometer Gravitational-Wave Observatory gravitational-wave detector
# Q18461808  Ligo — Italian town
# Q36946800  Ligo — family name
# Q30026643  Ligonchio — chief town of the homonym municipality
# Q30026645  Ligosullo — chief town of the homonym municipality
# Q18494373  Ligone (Santa Maria)
# Q2520468   Ligonier
# Q70330     Ligornetto
# Q785105    Ligonier, Pennsylvania — borough of Pennsylvania
# Q1781427   Ligorio López — Mexican footballer


  • -l, --lang <lang>: specify the results labels and descriptions language
  • -t, --limit <num>: set a custom limit (defaults to 10)
  • -j, --json: format the result as JSON
  • -v, --verbose: display rich results (aka summaries)

Verbose mode only:

  • -p, --properties <properties>: request additional properties (separated by a comma)

wd label

wd label <entities ids>
# Alias:
wd l <entities ids>
wd label Q1103345
# => The Cluetrain Manifesto
wd label Q18120925 Q22117436 Q22117437


  • -c, --clipboard: copy the command's result to the clipboard
  • -l, --lang <lang>: specify the label's language
wd label Q1103345 -l de
# => Cluetrain-Manifest
wd label Q123 -l zh
# => 9月

wd description

wd description <entities ids>
# Alias: (the d was already used by 'wd data')
wd desc <entities ids>
wd description Q1103345
wd description Q18120925 Q22117436 Q22117437


  • -c, --clipboard: copy the command's result to the clipboard
  • -l, --lang <lang>: specify the description's language

wd aliases

wd aliases <entities ids>
# Alias:
wd a <entities ids>
wd aliases Q1103345
wd aliases Q18120925 Q22117436 Q22117437


  • -c, --clipboard: copy the command's result to the clipboard
  • -l, --lang <lang>: specify the aliases's language

wd claims

A quick way to access the claims of an entity

wd claims <entities ids> [property id or pattern]
# Alias:
wd c <entities ids> [property id or pattern]
# all Q2001's claims
wd claims Q2001
# or just his place of birth
wd claims Q2001 P19
# or just the claims from properties of type Url
wd claims Q2001 Url
# or just the claims from properties with labels matching "library"
wd claims Q2001 library
# or website, etc
wd claims Q2001 website


  • -l, --lang <lang>: specify the properties labels' language
wd claims Q2001 -l es
wd claims Q2001 P19 --lang ru

Options when passing both an id an property:

  • -c, --clipboard: copy the command's result to clipboard
  • -j, --json: format the result as JSON

wd data

A quick way to access an entity's raw JSON data

wd data <entities ids>
# Alias:
wd d <entities ids>
wd data Q1496

This simply outputs the result of, parsed to keep only what is relevant to the requested entity (here Q1496). The output is valid json, so it lets you the possibility to pipe it to a JSON parser such as jsondepth:

wd data Q1496 | jd
# => { language: 'pt', value: 'Fernão de Magalhães' }

multiple entities

You can also request several entities at once by passing several ids. This outputs newline delimited JSON: one entity per-line, each line being valid JSON, but not the whole file as a whole.

wd data Q1496 Q123

Alternatively, you can pass ids from stdin:

echo "Q1496 Q123" | wd data

This especially make sense when you have thousands of ids to pass, passing them as arguments would fail. See wd sparql all-instances example.

simplified entities

You can request entities to be simplified, using wikidata-sdk simplify.entity function

wd data --simplify Q515168
# pass options to the simplify function
wd data --simplify --keep ids,references,qualifiers,hashes,nontruthy Q123
claims simplification keep options

You can customize the output by passing the list of data elements to keep among ids, references, qualifiers, hashes, nontruthy

wd data --simplify --keep ids,references,qualifiers,hashes,nontruthy Q123
# equivalent to
wd data --simplify --keep all Q123
# Can be useful to easily access claims ids
wd d -sk ids Q123 | jd .claims.P138 -j

See options for behavior details

filtered properties

Only request properties you need among labels,descriptions,aliases,claims,sitelinks

wd data --props labels,claims,sitelinks Q515168

wd revisions

Get entities revisions data

wd revisions <entities ids>
# Alias:
wd r <entities ids>
wd revisions Q3548931
wd r Q3548931


  • -s, --start <time>: specify a start time (date format)
  • -e, --end <time>: specify an end time (date format)
  • -l, --limit <time>: specify a limit number of revision (default and max = 500)

wd id

This one is kind of the inverse of wd label: pass it the title of a Wikipedia article and it will return the corresponding Wikidata id

wd id <article title or any Wikimedia project URL>
wd id Cantabria
# => Q3946
wd id New Delhi
# => Q987


  • -v, --verbose: make the output more verbose
  • -c, --clipboard: copy the command's result to the clipboard
  • -l, --lang <lang>: specify from which language the title comes By default, it will look at the Wikipedia corresponding to your environment local language (process.env.LANG), but you can specify another language by passing a 2-letters language code
wd id -l fr science politique
# => Q36442

You can also pass it full Wikipedia urls and let it deduce the language from there

wd id
# => Q9358
wd id
# => Q3816

wd props

A command to access the list of all Wikidata properties in a given language (by default the environment local language)

wd props [filter]
# Alias:
wd p [filter]

By default, your config or environment language is used:

wd props

Outputs a JSON object of the kind:

  "P2897": "identifiant Eldoblaje Movie",
  "P2898": null,
  "P2899": "âge minimal",

NB: properties without a label in the requested language are set to null, as you can see above for P2898 in French

This is especially convenient when you're looking for a property:

# look for a property having "image" in its label, description or aliases (case insensitive unless the argument contains capitals)
wd props photo


  "P18": "image",
  "P344": "director of photography",
  "P1259": "coordinates of the point of view",
  "P1442": "image of grave",
  "P1847": "Nasjonalbiblioteket photographer ID",
  "P1947": "Mapillary ID",
  "P1966": "Biblioteca Nacional de Chile catalogue number",
  "P2033": "Category for pictures taken with camera",
  "P2061": "aspect ratio",
  "P2485": "Fashion Model Directory photographer ID",
  "P2634": "model",
  "P2750": "Photographers' Identities Catalog ID",
  "P3269": " ID",
  "P3293": "BALaT image ID",
  "P4082": "image captured with",
  "P4640": "photosphere image",
  "P4759": "Luminous-Lint ID",
  "P5023": "activity policy in this place",
  "P5160": "Thesaurus For Graphic Materials ID"

Matching on labels, descriptions, and aliases can come very handy in cases such as the following:

wd props RSS


  "P1019": "feed URL"

But in case you want to match only on labels (ignoring descriptions, and aliases), you can pass the pattern to grep instead:

wd props | grep photo

Get the list of all Wikidata properties in another language

Option: -l, --lang <lang>: specify the properties labels language

wd props -l sv
# outputs the properties in Swedish

Get the list of all Wikidata properties with their types

Every property accepts values of a precise type, one of CommonsMedia, ExternalId, String, WikibaseItem, Time, GlobeCoordinate, Monolingualtext, Quantity, Url, WikibaseProperty, or Math

Option: -t, --type: include properties types

wd props --type


  "P6": { "label": "head of government", "type": "WikibaseItem" },
  "P10": { "label": "video", "type": "CommonsMedia" },
  "P14": { "label": "graphic symbol of thoroughfare", "type": "CommonsMedia" },
  "P15": { "label": "route map", "type": "CommonsMedia" },
  "P16": { "label": "highway system", "type": "WikibaseItem" },
  "P17": { "label": "country", "type": "WikibaseItem" },
  "P18": { "label": "image", "type": "CommonsMedia" },
  "P19": { "label": "place of birth", "type": "WikibaseItem" },
  "P20": { "label": "place of death", "type": "WikibaseItem" },

Get the list of all Wikidata properties of a given type

Re-using the possibility to pass a pattern to match, you can pass a property type

wd props --type Url
wd props --type CommonsMedia

NB: make sure to respect the case to get an exact match, otherwise it only match on the labels, descriptions and aliases.

Get the list of all Wikidata properties with their labels, types, descriptions, and aliases

wd props --details

Reset properties

wd props first tries to find the list in the props folder (created at wikidata-cli root), and request them to if missing.

This means that after a while, your local version will miss new and updated properties: this can be solved by using the --reset or --no-cache options:

Option: -n, --no-cache: ignore properties cache: fetch properties from the SPARQL endpoint, even if already cached Option: -r, --reset: clear cached properties, in all languages

wd sparql

A command to run a SPARQL query and get its JSON output

wd sparql <sparql or javascript file path>

static request from a SPARQL file

From this SPARQL query file:

# author_works.rq
  ?work wdt:P50 wd:Q42 .

get its output from your terminal like so:

wd sparql ./path/to/author_works.rq > ./results.json


  • -v, --verbose: log the generated SPARQL
  • -r, --raw: output raw SPARQL results (instead of results simplified by wikidata-sdk simplifySparqlResults function
  • -x, --index: get the results indexed by one of the SELECTed variables
  • -e, --sparql-endpoint <url>: customize the SPARQL endpoint (see below for examples)
wd sparql ./path/to/query.rq > ./results.json
wd sparql ./path/to/query.rq --raw > ./raw_sparql_results.json
wd sparql ./path/to/query.rq --index someVariableName > ./results_index.json

dynamic request from a JS file

Alernatively, you can pass the path from a javascript file exporting a function, the remaining arguments will be passed to the function:

// author_works.js
module.exports = authorId => {
  return `SELECT ?item WHERE {
    ?item wdt:P50 wd:${authorId} .
wd sparql ./path/to/author_works.js Q535 --json > ./Q535_works.json
wd sparql ./path/to/author_works.js Q5879 --json > ./Q5879_works.json
# This simple query could actually have been done using the `wd query` command
wd query --property P50 --object Q5879
# but, meh, let's keep it simple for the demo

You can use it to build alias commands for the requests you use often: the above can then be written

alias authors_works="wd sparql ./path/to/author_works.js --json"
authors_works Q535 > ./Q535_works.json
authors_works Q5879 > ./Q5879_works.json

Demo: Add book entities descriptions

wellknown queries

Some idiomatic queries than can't be done with wd query are included for convenience:


Fetch instances and instances of sub-classes of a given item. Example, all the instances of languages (Q34770):

wd sparql all-instances Q34770

This can be used to fetch the data associated with all the instances of a given item. For instance, to fetch all the painting's data, you could do:

# Get the ids of all the paintings
wd sparql all-instances Q3305213 > ./paintings_ids
# Fetch their data in a simplified format, and output it all as newline-delimted JSON, one entity per line (see wd-data for details)
wd data --simplify < ./paintings_ids > ./paintings.ndjson

custom SPARQL endpoint

The wd sparql command can actually be used with on other SPARQL endpoints:

wd sparql --sparql-endpoint bnf_request.rq
wd sparql --sparql-endpoint --raw --index item dbpedia_request.rq

wd query

wd query <options>
# Alias:
wd q <options>

A command to generate and run a simple SPARQL query, passing one or two of the elements that make a statement:

  • -s, --subject
  • -p, --property
  • -o, --object
# what is the entity id matching the twitter username (P2002) "timberners_lee"?
wd query --property P2002 --object timberners_lee
# which works have exoplanets (Q44559) for main subject (P921)?
wd query --property P921 --object Q44559 --labels
# or with the short options syntax
wd query -p P921 -o Q44559 -a

Other options:

  • -r, --raw: output raw SPARQL results (instead of results simplified by wikidata-sdk simplifySparqlResults function
  • -a, --labels: requests results labels
  • -l, --lang <lang>: specify the labels' language
  • -t, --limit <num>: set the request results limit (defaults to 1000)
  • -v, --verbose: log the generated request

wd open

wd open <entities ids>
# Alias:
wd o <entities ids>

A command to open a pages on Wikidata in a browser from the command line (yep, you can be that lazy)

open entities and properties pages

wd open Q123
# opens in your default browser

wd open P659
# opens

# also working with any string that matches /(Q|P)\d+/
wd open
# opens


  • -p, --wikipedia: open the Wikipedia article instead
  • -u, --url: simply generate the url, don't open it in a browser
wd open -p Q123
# opens because my system language is French
  • -l, --lang <lang>: specify which Wikipedia edition should be targeted
wd open -p -l sv Q123
# opens instead
wd open -p -l sv Q123 -u
# outputs without opening it in the browser

open a search page

wd open Dan Simmons
# opens