Spectra access

Gaia DR3 publishes for the first time spectra: - high resolution spectra around the Calcium triplet from the radial velocity spectrometer (RVS) - low resolution BP/RP spectra. DR3 provides about 33 million RVS spectra and more than 200 million BP/RP spectra.

In both cases the spectra are published in the form of arrays. This particularity leads to some restrictions discussed here.

Sampled vs. Continuous

The low resolution BP/RP spectra are published into two forms:

  • all spectra are published in the continuous form
  • a sub-selection of the published spectra are in pre-sampled form.

Querying_ pre-sampled_ spectra delivers flux and flux error for given wavelength. However, querying spectra in the continuous form provides a array of coefficients which can be used to reconstruct a sampled spectrum (see documentation).

Gaia Xp

DPAC provides a python module called GaiaXPy to reconstruct spectra from a continuous query. For more information refer to the offical documentation.

How to access the spectra

At Gaia@AIP we offer various ways to access spectra:

Direct download

When querying via the web interface, you can use the Download button to download as many spectra as you need at once. At Gaia@AIP there is no imposed limit!

Warning: Only .csv format is supported for spectra, the other formats will fail silently.

The .csv format is compatible with GaiaXPy, but can also be easily read with pandas: read-array-csv.py

Simple Join Service

For the users that wish to access the spectra in a scriptable manner, we offer a simple join service (see tutorial). This service can be used with PyVO and/or TOPCAT. It is a standard UWS service.

We also offer a web interface to access the spectra via the Simple Join Service.

The Simple Join Service was designed to be transparent for the users who are familliar with ESAC's Gaia Datalink Job service.

Warning: alike with ESAC's service there is a fixed maximum limit of 2000 spectra per query.


TAP does not support arrays. Querying a spectra table with TAP works, but the array type will be a char(32). Please use the simple join service instead.


Format Service Tables Arrays GaiaXPy
.csv download yes yes yes
.votable download yes no no
.ecsv simple-join-service yes yes no
.votable simple-join-service yes yes yes