Data Archive Server structure and contents

The Data Archive Server (DAS) is a file system containing all survey data products. It is accessible via

• http at http://das.sdss.org/DR4/data
• rsync at rsync://dr4@rsync.sdss.org/DR4 with password dr4.

This page mirrors the directory structure of the DAS, explaining which files are found where. It links to the "data model" descriptions of individual files. Directory names link to the http location, but the rsync server uses the same directory structure. For more details, see the data model for directories.

The root directory contains a number of ASCII parameter files with names fileList-*.par. These contain lists of all files in the archive with a CRC checksum for each, and of all files which are "missing" from the archive (i.e., you may expect to find these, but they do not exist).

Imaging data

• /DR4/data/imaging

The root directory for imaging. It contains two kinds of subdirectories: two inchunk directories containing the calibrated object lists, and one directory for each run/rerun pair containing the remaining imaging data.

inchunk directories

• /DR4/data/imaging/inchunk_best Download this directory to obtain the complete imaging catalogues with the best available data and reductions.
• /DR4/data/imaging/inchunk_target Download this directory to obtain the complete imaging catalogues as used to run target selection for the spectroscopic survey.

For either inchunk_best or inchunk_target, be sure to read the caveats on the calibrated object lists page.

The imaging is done along great circles. A unique covering of the sky is ensured by dividing these great circles into chunks. Calibrated object lists for each chunk are stored in subdirectories with names:

    stripeNN_muMMMMMM_S

E.g. /DR4/data/imaging/inchunk_best/stripe09_mu614999_1 .

Each of these stripeNN_muMMMMMM_S directories contain the following:

• Six sub directories (numbered 1 through 6), one per camera column, containing object lists as tsField*.fit and tsObj*.fit files.
• A mask subdirectory containing csv imaging masks. Note the csv masks are generated from the fpM files, which are stored in the corresponding run/rerun directory (see below).

Each best chunk stripeNN_muMMMMMM_1 directory contains the following files:

• tsChunk*.par: ASCII parameter file describing which imaging runs, reruns and fields cover this chunk.
• summary.php
• usno-N-MMMMMM.gif

Each target chunk stripeNN_muMMMMMM_0 directory contains the following files:

• tsChunk*.par: ASCII parameter file describing which imaging runs, reruns and fields cover this chunk.
• tsTargets*.fit: Lists all spectroscopic targets selected from a given chunk.

run/rerun directories

These directories have names of the form

• /DR4/data/imaging/RRRR/EE/

where RRRR is the run number and EE is the rerun number, eg. /DR4/data/imaging/1336/40 .

Each run/rerun directory contains the two types of subdirectories.

• Directories which contain only files:
  • astrom -- results of the astrom pipeline
  • logs -- opBC*.par files listing bad columns for this imaging run
  • nfcalib -- results of photometric calibration
• Directories with one subdirectory per camera column (1-6), containing the bulk of the imaging data:
  • corr -- corrected frames fpC*.fit.gz, ie. reduced survey images
  • objcs -- further outputs of photometric pipeline:
    • fpAtlas*.fit -- set of pixels belonging to each object in the calibrated object list (corresponding tsObj file). See how to read an atlas image
    • fpBIN*.fit -- 2x2 and 4x4 binned versions of corrected frames (fpC) after objects detected on original frame have been subtracted
    • fpM*.fit -- full masks with saturated pixels etc. See how to read fpM masks. These form the basis for the csv imaging masks (see above).
    • fpObjc*.fit -- object lists in counts and pixels (these become the tsObj files after photometric calibration and conversion of scales in pixels to arcseconds, among others)
    • fpFieldStat*.fit -- statistics of photometric pipeline processing for a single field. Basis for tsField files.
    • psField*.fit -- Prelimary photometric calibration, as well as final point-spread-function fit, for a single field. Use these to reconstruct the PSF at any position.
  • Zoom -- fpC*.jpeg images of the sky in various zoom levels

Spectroscopic data

• /DR4/data/spectro/2d_23/

spectro2d pipeline output, produced by rerun 23 of the spectro pipelines: the calibrated spectra. Contains one directory per plate with name:

• /DR2/data/spectro/2d_23/PPPP

where PPPP is the plate number (eg. /DR4/data/spectro/2d_23/0644). Each directory contains one spPlate*.fit file with all 640 calibrated spectra for that plate.

• /DR4/data/spectro/1d_23

Calibrated spectra (including gif and ps plots) and classifications, redshifts, and other measured parameters (line fits, velocity dispersions, line indices). Contains one directory per plate with name:

• /DR4/data/spectro/1d_23/PPPP

where PPPP is the plate number (eg. /DR4/data/spectro/1d_23/0644).

Each plate directory contains two subdirectories:

• 1d -- contains 640 spSpec*.fit files, one per spectrum, containing the spectrum and all derived parameters.
• gif -- contains spPlot*.gif, and spPlot*.ps.gz files for each spectrum, with line annotations

Spectra + imaging

• /DR4/data/spectro/ss_23
• /DR4/data/spectro/ss_tar_23

See separate page about getting spectra in bulk.