2dFGRS Final Release Database
The 2dFGRS database consists of three main components:
There are 382323 target objects in the final FITS database, each with its own FITS file. Each object in the survey source catalogue has been given a serial number (SEQNUM), and the name of the FITS file for that object is this serial number. The serial numbers for objects in the SGP strip are 1-193550, for objects in the NGP strip 193551-332694, and for objects in the random fields 332695-389713.
Note that: (i) The 'supplementary' objects with serial numbers >389713 that were included in the 100k release have now all been deleted; these were additional objects fainter than the nominal survey limit, and were never observed. (ii) The objects in random field 859 were improperly named, so, as this field was never observed, the 1166 sources in the field have been deleted. (iii) Visual inspection of the source images during the initial construction of the source catalogue identified 4604 sources as artifacts; these have now been deleted. (iv) A final visual check of problematic objects selected by comparing the APM and SuperCosmos source lists (see below) identified a further 1620 sources as artifacts; these have also been deleted. These deletions mean that the source catalogue in the final release is significantly cleaner than that in the 100k release.
Each FITS file has a primary part which contains all the source catalogue data about the object (as FITS keywords) and a Digital Sky Survey sky chart (postage stamp image) of the objects. The WWW version of the database also contains SuperCosmos bJ and rF postage stamp images (these images are omitted from the DVD version to conserve space, although the SuperCosmos parameters are present in both versions). Subsequent extensions of the FITS file contain the spectral observations and parameters obtained for the object, appended in chronological order. Each spectrum extension contains the spectrum of the object, the variance (error) array for the object spectrum, the spectrum of the mean sky that was subtracted from the object spectrum, and FITS keyword data giving information about the spectroscopic observation and derived parameters such as the redshift and spectral quality. Many targets contain multiple spectrum extensions corresponding to multiple observations.
Table 1 lists all the FITS keywords present in the primary part (extension 0) of the FITS files; Table 2 lists all the keywords present in spectrum extensions 1..spectra. The tables give the names of the keywords, example values, and the keyword descriptions.
The first set of photometric parameters are from APM scans of the Southern Sky Survey bJ plates, and are the primary photometric data on which the 2dFGRS was based. Photometric parameters with FITS keywords prefixed SB and SR are from SuperCosmos scans of, respectively, the bJ and rF Southern Sky Survey plates.
In order to remove artifacts from the source catalogue and to flag other problematic cases, 10812 sources (about 3% of the total) were selected for visual examination of their images based on various criteria: 1656 objects with no SuperCosmos source matching the APM position, 4377 objects with a poor match between the APM and SuperCosmos positions and/or magnitudes, another 1321 objects with anomalous colours, and finally all 3458 bright objects with SuperCosmos magnitude bJ<16. These objects' images were individually examined and the VISCHECK classification parameter set as follows: 0 = not examined (373131 sources), 1 = galaxy (5283 sources), 2 = star (114 sources), 3 = merger (3102 sources), 4 = in bright star halo (377 sources), 5 = artifact (1620 sources), 6 = other (316 sources). In addition, the VISCOMM parameter gives comments in some cases (and is mandatory for 'other' classifications). The 1620 objects classified as artifacts were deleted from the database; all 9192 other sources are retained.
Table 1: FITS keywords for the primary image (extension 0) Keyword = Value / Description SIMPLE x T / file does conform to FITS standard BITPIX x 16 / number of bits per data pixel NAXIS x 2 / number of data axes NAXIS1 x 49 / length of data axis 1 NAXIS2 x 49 / length of data axis 2 EXTEND x T / FITS dataset may contain extensions BSCALE x 1.0000 / REAL = (FITS * BSCALE) + BZERO BZERO x 0.0000 / Bias SEQNUM = 100100 / Sequence number : Database Primary Key NAME = 'TGS469Z164' / 2dFgg assigned name IMAGE x 'SKYCHART' / Existance of postage stamp image RA = 0.7943429758 / RA in radians : 3 2 3.00 DEC = -0.5475286941 / Dec in radians : -31 22 15.9 APMEQNX = 1950.00 / Equinox of RA and Dec BJSEL = 18.858 / final bj mag used in the object selection PROB = 2335.4 / psi classification parameter PARK = 0.910 / k classification parameter PARMU = 0.187 / mu classification parameter IGAL = 1 / final class flag (equals 1 for a galaxy) JON = -1 / eyeball class flag ORIENT = 91.0 / orientation in degrees clockwise from E-W ECCENT = 0.270 / eccentricity AREA = 308.0 / isophotal area in pixels X_BJ x 2918.7 / plate x_bj in 8 micron pixels Y_BJ x 9123.1 / plate y_bj in 8 micron pixels DX x 43.0 / corrected difference (x_bj - x_R)*100 DY x 49.0 / corrected difference (y_bj - y_R)*100 BJG = 18.920 / bj without extinction correction RMAG x 10.35 / unmatched apm "total" mag PMAG x 10.53 / unmatched raw apm profile integrated mag FMAG x 8.72 / unmatched raw apm 2" profile integrated mag SMAG x 10.74 / unmatched raw stellar mag (from APMCAL) REDMAG x / unmatched raw red stellar mag (from APMCAL) IFIELD = 417 / ukst field IGFIELD x 2007 / galaxy number inthis field REGION = 'S417 ' / GSSS Region name OBJEQNX = 2000.00 / Equinox of the plate reference frame OBJRA = 0.8034094522 / RA in Radians : 03 04 07.673 OBJDEC = -0.5441390223 / Dec in Radians : -31 10 36.73 PLTSCALE x 67.2000 / Plate Scale in arcsec per mm XPIXELSZ x 25.2844500 / X pixel size in microns YPIXELSZ x 25.2844500 / Y pixel size in microns OBJPLTX x 7970.86 / Object X on plate (pixels) OBJPLTY x 4148.11 / Object Y on plate (pixels) DATAMAX x 14431 / Maximum data value DATAMIN x 4011 / Minimum data value BJSELOLD = 18.96 / original bj mag used in the object selection BJG_OLD = 19.01 / original bj without extinction correction BJSEL100 = 18.93 / bj mag used in the object selection (100k) BJG_100 = 18.99 / bj without extinction correction (100k) GALEXT = 0.062 / galactic extinction VISCHECK = 0 / visual check flag VISCOMM = ' ' / visual check comments SBFIELD x '417 ' / Sky survey field number SBRA = 46.032062700000 / Object RA (J2000.0) SBDEC = -31.176815800000 / Object declination (J2000.0) SBBJMAG = 18.836 / UK-J (Bj) magnitude SBAREA = 254 / Total area SBAI = 7927 / Weighted semi-major axis SBBI = 6103 / Weighted semi-minor axis SBPA = 8 / Celestial position angle SBCLASS = 1 / Classification flag SBSIGMA = 22.668 / profile classification statistic SBBLEND = 0 / Deblending flag SBQUAL = 0 / Quality flag SBAPM = 1 / APM-Supercosmos Bj-band match classification SRFIELD x '417 ' / Sky survey field number SRRA = 46.031963000000 / Object RA (J2000.0) SRDEC = -31.176771700000 / Object declination (J2000.0) SRRMAG = 17.188 / UK-J (R) magnitude SRAREA = 249 / Total area SRAI = 7467 / Weighted semi-major axis SRBI = 6317 / Weighted semi-minor axis SRPA = 13 / Celestial position angle SRCLASS = 1 / Classification flag SRSIGMA = 16.574 / profile classification statistic SRBLEND = 0 / Deblending flag SRQUAL = 0 / Quality flag SRSB = 1 / Supercosmos R vrs Bj match classification
Table 2: FITS keywords for the spectra (extensions 1..spectra) Keyword = Value / Description XTENSION x 'IMAGE ' / IMAGE extension BITPIX x -32 / number of bits per data pixel NAXIS x 2 / number of data axes NAXIS1 x 1024 / length of data axis 1 NAXIS2 x 3 / length of data axis 2 PCOUNT x 0 / required keyword; must = 0 GCOUNT x 1 / required keyword; must = 1 CRVAL1 x 5802.8979492 / Co-ordinate value of axis 1 CDELT1 x 4.3103027344 / Co-ordinate increment along axis 1 CRPIX1 x 512.0000000000 / Reference pixel along axis 1 CUNIT1 x 'Angstroms' / Units for axis 1 EXTNAME x 'SPECTRUM' / 2dFGRS spectrum OBSNAME = 'TGS469Z164' / Observed object name OBSRA = 0.7943429758 / Observed RA OBSDEC = -0.5475286940 / Observed Dec MATCH_DR = 0.0000 / Position match dr in arcsec Z = 0.178876 / raw measured redshift Z_HELIO = 0.178860 / heliocentric corrected redshift QUALITY = 5 / redshift measurement quality ABEMMA = 1 / redshift source abs1 emm2 man3 NMBEST = 0 / number emission lines for best z_em NGOOD = 0 / number of good emmision lines Z_EMI = -9.9990 / emission redshift Q_Z_EMI = 0 / emission redshift quality KBESTR = 2 / x-correlation template R_CRCOR = 15.5600 / x-correlation peak Z_ABS = 0.1789 / x-correlation redshift Q_Z_ABS = 3 / x-correlation quality Q_FINAL = 3 / Suggested quality for redshift IALTER = 0 / No idea Z_COMM = ' ' / Observers Comment THPUT = 0.96613 / Fibre Throughput SPFILE = 'sgp469_991104_1z.fits' / 2dF reduced data file PLATE = 1 / 2dF plate number PIVOT = 302 / 2dF pivot number FIBRE = 58 / 2dF fibre number OBSRUN = '99OCT ' / Observation run GRS_DATE = '991104 ' / 2dF YYMMDD observed date UTDATE = '1999:11:04' / UT date of observation SPECTID = 'A ' / 2dF spectrograph ID GRATID = '300B ' / 2dF grating ID GRATLPMM = 300 / 2dF grating line per mm GRATBLAZ = 'COLLIMATOR' / 2dF grating blaze direction GRATANGL = 25.30000 / 2dF grating angle LAMBDAC = 5782.700 / Central wavelength CCD x 'TEKTRONIX_5' / CCD ID CCDGAIN x 2.790 / CCD inverse gain e/ADU CCDNOISE x 5.200 / CCD read noise (electrons) OBJX = 196833 / 2dF object x position OBJY = 10401 / 2dF object y position OBJXERR = 6 / 2dF object x position error OBJYERR = 14 / 2dF object y position error OBJMAG = 18.96 / 2dF object magnitude THETA = 4.526 / 2dF fibre angle PTRTYPE = 'P ' / 2dF ptrtype PID = 0 / 2dF pid OBSFLD = 'sgp469 ' / 2dF observed field number NCOMB = 3 / Number of frames combined REFRUN = 31 / AAT Run number of reference frame UTSTART x '16:37:59.48' / UT start of exposure of reference run UTEND x '16:57:59' / UT end of exposure of reference run REFEXP = 1200.0 / Exposure time (secs) of reference run REFHASTA x '36.07263999999999982' / HA at start of exposure of reference run REFHAEND x '41.08118999999999943' / HA at end of exposure of reference run SNR = 2.0299999E+01 / Median signal-to-noise ratio per pixel ETA_TYPE = -2.5934000E+00 / Eta spectral type (2dF defined) SEEING = 3.0000000E+00 / Seeing (2dF calculated)
The mSQL database (see Jepson & Hughes 1998) can be thought of as a table. The rows of the table are labelled by the unique object serial number (serial, identical to the parameter SEQNUM in the primary extension of the FITS file) and the extension number (extnum). There are multiple rows for each target object corresponding to each of the extensions in the object's FITS file: the first row corresponds to the primary FITS extension (extension 0), while subsequent rows correspond to the spectrum extensions 1..spectra. The columns of the table correspond to the object parameters, and are labelled by the name of the corresponding keyword. N.B. case is significant in these keywords.
The object serial numbers (serial) provide the primary database key, but the objects are also indexed by their unique survey name (name, identical to the parameter NAME in the primary extension of the FITS file), which has the format TGhfffZnnn, where h is the hemisphere (N for the NGP strip and S for the SGP strip and random fields), fff is the number of the primary field to which the object is assigned and nnn is the number of the galaxy within that field.
Note that the observed name of the object (parameter OBSNAME in each spectrum extension) is the same as name (or NAME in the primary extension of the FITS file) except that: (i) if the field in which the object is observed (given by OBSFLD) is an overlapping field rather than its primary field (given by fff), then the first character of the name is changed from T to X; and (ii) if the object has been flagged as a possible merger, then the second character of the name is changed from G to M.
The first row for each object (extnum=0) contains the source catalogue data and the basic spectroscopic information for the best spectrum of that object. The keywords for each row are a subset of the FITS parameters for the primary image (the parameters in Table 1 with an '=' between Keyword and Value; those with an 'x' instead of an '=' are not in the mSQL database) plus all the additional keywords listed in Table 3. The best spectrum is the one with the highest redshift quality parameter; if there is more than one spectrum of the same quality, then the latest of these spectra is taken to be the best.
Subsequent rows for the same object (extnum=1..spectra, where spectra is the number of spectra obtained for that object) contain a subset of the FITS parameters pertaining to each spectroscopic observation (the parameters in Table 2 with an '=' between Keyword and Value; those with an 'x' instead of an '=' are not in the mSQL database) plus the additional keywords in section (i) of Table 3. If there is no spectrum for the object then spectra=0 and only the row corresponding to extnum=0 will exist. Note that some information is duplicated between rows and that not all parameters are defined for all rows; undefined parameters return a NULL value.
The best spectrum is indicated by the parameter best, which is present in all extensions: in the primary extension (extnum=0) its value is the number of the extension with the best spectrum; in the spectral extensions (extnum>0) its value is 1 if that extension contains the best spectrum and 0 otherwise.
Table 3: Additional mSQL database keywords (i) Keywords in all extensions Keyword = Value / Description serial = 100100 / 2dFGRS serial number name = TGS469Z164 / 2dFGRS name UKST = 417 / UKST sky survey field number spectra = 1 / number of spectra for this object extnum = 1 / extension number best = 1 / in extension 0, extension number with best spectrum / in extension>0, 1 if best spectrum and 0 otherwise obsrun = 99OCT / observing run year and month TDFgg = -469 / 2dFGRS field number (+NGP,-SGP) pivot = 302 / 2dF pivot plate = 1 / 2dF plate fiber = 58 / 2dF fiber z = 0.178876 / observed redshift z_helio = 0.178860 / heliocentric redshift abemma = 1 / redshift type (abs=1,emi=2,man=3) quality = 5 / redshift quality parameter (ii) Keywords in extension 0 only Keyword = Value / Description alpha = 0.7943429758 / RA (B1950) in radians delta = -0.5475286941 / DEC (B1950) in radians ra = 3 2 3.00 / RA (B1950) in HH MM SS.SS dec = -31 22 15.9 / DEC (B1950) in DD MM SS.S ra2000 = 03 04 07.68 / RA (J2000) in HH MM SS.SS dec2000 = -31 10 36.8 / DEC (J2000) in DD MM SS.S l2 = 228.9258834424 / Galactic longitude b2 = -60.8572447739 / Galactic latitude
Searches of the database use the mSQL query format (Jepson & Hughes,
1998), which has the general format
An example is
Another example with a more complex list of conditions is
One common and useful type of search is where parameters from only
the best spectrum (or from the best spectrum and the primary extension)
are desired. This can easily be achieved using the best parameter. For example,
Simple searches on the two indexed parameters (serial and name), are quick - e.g. WHERE serial=69656 or WHERE name='TGS203Z081'; more complex searches take about 5 minutes. Use of the ORDER BY construction is very slow for large datasets, and is not generally recommended.
Further information about the mSQL database software and its
structured query language is given in Yarger et al. (1999) and on the
The 2dFGRS mSQL database can be searched via the WWW interface in a number of ways:
The results of a query can be returned either directly as an HTML table (suitable for relatively small datasets) or via an email giving the URL of the results file (suitable for large datasets). The results file may be either a compressed text (gzipped ASCII) file containing the chosen parameters for the objects selected by the mSQL query, or a compressed archive (gzipped tarfile) of the FITS files for the objects selected by the query, depending upon the option selected. If results are returned as an HTML table, then objects can be selected interactively and their DSS images (and for the on-line database, their SuperCosmos images) and their spectra can be displayed. If the spectra have measured redshifts, then the plot of each spectrum shows the positions of prominent spectral features at the redshift associated with that spectrum.