Common attributes

A handful of attributes appear on nearly every OpenAlex entity. They mean the same thing everywhere, so they’re documented here once; each entity’s own attribute dictionary links back to this page rather than repeating them.

Identity

id

String. The canonical OpenAlex ID for this entity, as a URL, e.g. https://openalex.org/A5023888391. Stable and never reused; if the entity is later merged into another, requests for this ID usually return the survivor (best effort, not a permanence promise).

ids

Object. All known identifiers for this entity, expressed as URIs where possible, keyed by namespace — always openalex, plus whichever external IDs apply to the entity type (doi, orcid, ror, issn_l, wikidata, mag, …). A key with no known value is left out on some entities (works, topics) and present as null on others (authors, sources, institutions, publishers, funders, concepts), so check for both.

display_name

String or null. The entity’s name — a work’s title, an author’s name, a source’s title, an institution’s name. Every entity has one, except a work or award that no source gives a title, where it is null.

Scale & influence

works_count

Integer. How many works are associated with this entity — authored by an author, published in a source, affiliated with an institution, tagged with a topic, and so on. For works themselves this field is absent. It counts pulled works (retracted, withdrawn or removed), which a works list leaves out by default, so the list can come out shorter.

cited_by_count

Integer. The total number of citations this entity has received — summed across all of its works for non-work entities. Recomputed as the graph updates, so it drifts upward over time.

summary_stats

Object. A few precomputed bibliometric indicators for the entity:

  • 2yr_mean_citedness — a journal impact factor-style 2-year mean citedness: the citations received in the last full calendar year by the entity’s works published in the two years before it, divided by the number of those works. In 2026 that is the 2025 citations to works published in 2023 and 2024, read from each work’s counts_by_year; the window moves on 1 January. Every work counts (sources and publishers leave out peer reviews), so unlike the Journal Impact Factor the denominator is not limited to articles and reviews.
  • h_index — the h-index.
  • i10_index — the number of works with at least 10 citations.

These are convenience metrics; for anything rigorous, compute over the underlying works. Not available on works.

counts_by_year

List. Per-year rollup of the entity’s activity, each element { year, works_count, cited_by_count } (author, source, institution and funder objects also carry oa_works_count). Handy for sparklines and trend lines.

  • works_count (and oa_works_count) counts the entity’s works published in that year.
  • cited_by_count counts the citations the entity’s works received in that year, whenever those works were published. It is the sum of the works’ own counts_by_year, so it means the same thing as on a work. A year in which the entity was cited but published nothing appears with works_count: 0.
  • Citations are binned by year from 2012, so earlier years show cited_by_count: 0. Years after the current one never appear, and years with neither works nor citations are left out.
  • How far back the list goes depends on the entity type: every publication year for authors, sources and funders; from 2010 for institutions; from 2014 for publishers.

The array’s sort direction is not consistent across entity types — always sort by year yourself before slicing the most-recent N.

Provenance timestamps

created_date

String. When this entity was first added to OpenAlex, as an ISO 8601 date and time, e.g. 2016-06-24T00:00:00. Works, authors and sources record only the day, so their time is always 00:00:00; awards add milliseconds and a Z (2026-07-13T04:19:57.000Z). The first 10 characters are the date. A work’s created_date never moves later, even when its record is rebuilt (works).

updated_date

String. The datetime this entity’s record last changed (an ISO 8601 datetime). It moves for routine changes too, such as a new citation, not only for metadata edits.

Two exceptions, if you keep a copy current by updated_date (snapshot partitions or from_updated_date):

  • On works, four fields change without moving it: fwci, the value of citation_normalized_percentile and cited_by_percentile_year are recomputed every night for every work (a large share shift a little each time, so moving updated_date for them would send you most of the corpus every day), and has_content follows the files OpenAlex holds. A recent full snapshot has each night’s exact values; the API’s are within 1% of them (how fresh). The percentile’s is_in_top_10_percent and is_in_top_1_percent flags do move it (from 30 October 2026).
  • A change to every record at once can leave it unmoved, so that syncing customers don’t re-download the whole database for one field. Each such change is announced with what to reload, in the snapshot’s RELEASE_NOTES.txt or on the field’s own page: in October 2026, work keywords, sustainable_development_goals, related_works and created_date.

In the API

These fields are selectable, and most are filterable and sortable, on every entity endpoint — see Selecting fields and Filtering. The exact operations each field supports are listed in that entity’s attribute dictionary.

Last updated

View as Markdown