How many planets orbit other stars, and how were they found?

Steps

1. No key. One TAP request: select pl_name, hostname, discoverymethod, disc_year, disc_facility, sy_snum, sy_pnum, sy_dist, ra, dec, pl_rade, pl_bmasse, pl_orbper, disc_pubdate from pscomppars order by pl_name. 6,445 rows. The script grouped the rows by disc_year and discoverymethod, grouped them by host star for distances, and recomputed the total (6,445), the count for each of 11 discovery methods and the counts for Kepler, K2 and TESS; all equal the numbers on the archive's own statistics page (docs/counts_detail.html). For every host star, all its planets carry the same distance and position, and no host has more planet rows than the archive's planet count for its system (sy_pnum).
2. A confirmed planet is a row of the Planetary Systems Composite Parameters table (one row per planet, with values combined from different references). The archive includes a planet when its properties are in a refereed publication and its planetary status is unambiguous. Discovery year and method are the archive's own columns (disc_year, discoverymethod). Methods other than transit, radial velocity, microlensing and direct imaging are grouped as Other.
3. Distance: the archive's sy_dist in parsecs for the host star, converted at 3.26156 light years per parsec (648,000 / pi astronomical units per parsec; 63,241.077 astronomical units per light year).

35 yearly rows (1992 to 2026) from 6,445 planets. Unit: Number of confirmed planets; distance in light years. Every derived figure on this page (changes, indexes, averages, counts) is computed from the published rows, and the table under each chart says how.

Checks

• The total (6,445), the count for each of the 11 discovery methods and the counts for Kepler, K2 and TESS equal the numbers on the archive's own Exoplanet and Candidate Statistics page. The script stops if one differs.
• The yearly counts add up to the total; every planet in a host star's rows carries the same distance and position; no host has more rows than its system's planet count.
• The nearest host star is Proxima Centauri at 4.24 light years; the script stops if that changes, as it would signal a change in the table.
• The nearby-stars file (exo_nearby.json) holds the 100 nearest hosts, out to 41.47 light years, with right ascension and declination in degrees.

Caveats

• Discoveries are shaped by how each method works, so the counts are not a census of planets in the galaxy. Transit surveys find planets that cross their star as seen from here, and the Kepler and TESS missions looked at particular patches of sky, so the distance bands show where telescopes looked, not where planets are. Most planets in the archive are large enough to see; small, cool, distant planets are mostly missing. The latest year is partial. Discovery year is the year of the publication that announced the planet. A few values change when a paper revises them: the table is updated weekly. Distances are those of the host star in the archive, from parallax where it exists.
• 28 host stars have no distance in the archive and are left out of the distance charts. Discovery year is the year of the announcing paper.
• The archive counts a planet only when its properties are in a refereed publication and its planetary nature is unambiguous. Candidates (such as the thousands from TESS and Kepler that are not yet confirmed) are not counted.

Releases and revisions

• Table read through TAP on 2026-10-08; the most recent discovery publication date in the table is 2026-09
• The archive updates weekly. Next: re-run space_exoplanets.py to refresh; it makes two requests (the table and the statistics page).