skycell: a cost-based HEALPix index for sky positions in PostgreSQL, with native support for IVOA standards

Astronomical archives answer positional queries from relational databases, and in PostgreSQL two extensions dominate: Q3C, an integer key in a B-tree, and pgSphere, spherical geometry in a GiST tree. We present skycell, a PostgreSQL extension that keeps the integer key and the B-tree yet answers cones, polygons, cross-matches, nearest neighbours and stored footprints through the IVOA standards an archive already serves: ADQL’s geometry functions under their own names, so a TAP service passes queries through untranslated; STC-S regions, unions, intersections, differences and complements included, indexed in ObsCore s region columns; MOC 2.0 coverages in and out; and the positional functions of the IVOA catalogue of ADQL user-defined functions. Positions are keyed by their order-29 HEALPix cell, the numbering MOCs use, and a query region is covered by cells whose depth is chosen by a cost model: the price of an extra index range against the false-positive rows it removes, with local source density read from the histogram PostgreSQL already keeps. A planner custom scan walks the covering’s ranges, a GiST operator class indexes stored regions, and coverings are validated by brute force over millions of points. In randomized paired trials on PostgreSQL 18.6 with 10- and 50-million-row catalogues, including real Gaia DR3 positions, skycell is faster than pgSphere at every cone radius from 1′′ to 3◦, by 5–34% with a warm cache and up to 2.5 times cold. It cross-matches 16–27% faster than Q3C’s q3c join and three times faster than pgSphere, and answers convex polygons 4.3–4.8 times faster than pgSphere, from an index a third the size. One index serving points, regions and cross-matches through the IVOA standards makes skycell a strong default for an archive’s positional layer.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-15
DOI
https://doi.org/10.5281/zenodo.24236317
Primary Topic
Data Management and Algorithms
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

skycell: a cost-based HEALPix index for sky positions in PostgreSQL, with native support for IVOA standards

Michele Delli Veneri, Jesús J. Salgado
Zenodo (CERN European Organization for Nuclear Research)
Data Management and Algorithms
article

skycell: a cost-based HEALPix index for sky positions in PostgreSQL, with native support for IVOA standards

Michele Delli Veneri, Jesús J. Salgado
article en

Abstract

Astronomical archives answer positional queries from relational databases, and in PostgreSQL two extensions dominate: Q3C, an integer key in a B-tree, and pgSphere, spherical geometry in a GiST tree. We present skycell, a PostgreSQL extension that keeps the integer key and the B-tree yet answers cones, polygons, cross-matches, nearest neighbours and stored footprints through the IVOA standards an archive already serves: ADQL’s geometry functions under their own names, so a TAP service passes queries through untranslated; STC-S regions, unions, intersections, differences and complements included, indexed in ObsCore s region columns; MOC 2.0 coverages in and out; and the positional functions of the IVOA catalogue of ADQL user-defined functions. Positions are keyed by their order-29 HEALPix cell, the numbering MOCs use, and a query region is covered by cells whose depth is chosen by a cost model: the price of an extra index range against the false-positive rows it removes, with local source density read from the histogram PostgreSQL already keeps. A planner custom scan walks the covering’s ranges, a GiST operator class indexes stored regions, and coverings are validated by brute force over millions of points. In randomized paired trials on PostgreSQL 18.6 with 10- and 50-million-row catalogues, including real Gaia DR3 positions, skycell is faster than pgSphere at every cone radius from 1′′ to 3◦, by 5–34% with a warm cache and up to 2.5 times cold. It cross-matches 16–27% faster than Q3C’s q3c join and three times faster than pgSphere, and answers convex polygons 4.3–4.8 times faster than pgSphere, from an index a third the size. One index serving points, regions and cross-matches through the IVOA standards makes skycell a strong default for an archive’s positional layer.

Zenodo (CERN European Organization for Nuclear Research)
SKA Observatory (GB), Istituto Nazionale di Fisica Nucleare, Sezione di Napoli (IT)
Data Management and Algorithms
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon
skycell: a cost-based HEALPix index for sky positions in PostgreSQL, with native support for IVOA standards — Michele Delli Veneri, Jesús J. Salgado · Zenodo (CERN European Organization for Nuclear Research) (2026) | TGRS Research Map | TGRS