THE EUVE SATELLITE SURVEY DATABASE
            N. Craig, T. Chen, I. Hawkins, and A. Fruscione
              Center for EUV Astrophysics, 2150 Kittredge,
     University of California, Berkeley, California 94720, USA


ABSTRACT

   The EUVE survey database contains fundamental science data for 9000 po-
tential source locations (pigeonholes) in the sky.  The first release of the
Bright Source List is now available to the public through an interface with
the Astrophysical Data System of NASA.  We describe the database schema design
and the EUVE source categorization algorithm that compares sources to the ROSAT
Wide Field Camera source list.


INTRODUCTION

   The Extreme Ultraviolet Explorer Survey Database (EUVECATDB) was designed
and implemented to store and maintain all fundamental science data of sources
from the EUVE all-sky survey.  The survey phase, which lasted six months, was
completed January 21, 1993.  Information on the source position and flux in
each filter are calculated with the End-to-End System (EES) software [1].
These data are then stored in EUVECATDB, which is updated as new survey data
become available and reprocessed.  Currently, EUVECATDB contains about 9000
source locations.
   The database is implemented as a combination of a Sybase Transact Sequential
Query Language script and Unix shell scripts; procedures and triggers internal
to Sybase are applied to ensure data integrity at all times.  The six major
tables and their relationship within the database are shown in fig. 1.  Some
of the database's built-in algorithms were designed to carry out computations
automatically and also to categorize all entries based on a comparison with
sources detected by ROSAT's Wide Field Camera (WFC).  EUVECATDB will be in-
stalled in NASA's Astrophysics Data System (ADS) so that it is available to
the scientific community (after the data proprietary periods end) via a variety
of methods, including an  X-window user interface.  Currently the data from the
EUVE Bright Source List, a sample of the 356 brightest EUVE sources [2], are
installed in ADS and are accessible only by staff at the Center for EUV
Astrophysics.


2. DATABASE SCHEMA DESIGN

   The EUVECATDB  consists of six tables (fig. 1).  The pig table (pigeonhole
[3]), the backbone of the whole database, holds astronomical data for the ex-
pected targets,	such as spectral class, pixel address, RA and DEC.  Each data
row in this table corresponds to a pigeonhole (target) entry.  A unique integer
defined as pid, pigeonhole id, is assigned to each pigeonhole and uniquely
identifies the pigeonholes in all tables.
   Each data row in the fil (filter) table contains the observed data from a
detector's specific  filter when the detector was pointing to a pigeonhole.
Because EUVE has detectors for three scanning telescopes, with each detector
having multiple filters, each pigeonhole entry in the pig table has between
two and ten corresponding data rows in the fil table.  Each data row in the fil
table is uniquely identified by the combination of a pid and filter number.
   Each data row in the wfc table corresponds to a source detected by the ROSAT
mission and is uniquely identified by the attribute wfc_name.  Each data row in
the pig_wfc table consists of a pigeonhole id and a WFC source name that over-
lap.  Error circle radii within 90% confidence level, r90, are calculated both
for EUVE and WFC sources.  When the distance between EUVE and WFC source is less
than the sum of the two r90 values, these two sources are considered overlap-
ping.  EUVECATDB uses the data in the wfc and pig_wfc tables to carry out the
EUVE source categorizing algorithm.
   For each pigeonhole entry in the pig table, there is at least one corres-
ponding entry in the cat (catalog) table which stores a source name from a
different catalog.  Lastly, the candi (optical candidates) table holds the
data of potential optical counterparts.  For each pigeonhole there may be
more than one candidate entry that is considered a potential counterpart of
the EUVE source.  Each data row in the candi table is uniquely identified by
the combination of pid and cand_id.
   The following is an example of querying the database in the Sybase ISQL
environment.
   List the EUVE sources that overlap with WFC, their spectral types, and count
rates from EUVE and WFC.

	> select cand_id,wfc_name,sp_class,cnt_rate,s1
	> from pig_wfc, candi,fil
	> where pig_wfc.pid = candi.pid and candi.pid = fil.pid
	> go

	cand_id   wfc_name   sp_class  cnt_rt    s1a
	UXFOR     RE0243375     G5G     0.039     21
	VYARI     RE0248+310    G9e     0.056    109
	HD18131   RE0254051     K0      0.090     60
	EFERI     RE0314223     CVAM    0.248    181


3. CATEGORIZING PIGEONHOLES WITH RESPECT TO WRC SOURCES

   The Sybase procedure detect was implemented within EUVECATDB to automati-
cally carry out the categorization of EUVE sources by comparing them to WFC
sources.  This procedure, when executed, assigns an integer number to each
pigeonhole entry in the pig table to indicate the category to which the pigeon-
hole belongs.  There are five categories for the EUVE sources in comparison to
the WFC sources:

 (1) A new source:  the pigeonhole has a "detection" in one of the EUVE band-
	passes and does not overlap with any of the WFC sources.
 (2) WFC yes, EUVE more:  the pigeonhole has a "detection" in the Lexan band-
	pass (which corresponds to S1 from WFC) in addition to at least one
	other EUVE bandpass, and it overlaps a WFC source.
 (3) Common:  the pigeonhole has a "detection" in the EUVE Lexan bandpass only,
	and overlaps with a WFC source.
 (4) WFC yes, EUVE no:  the pigeonhole does not have a "detection" in any of
	the EUVE bandpasses but overlaps with a WFC source.
 (5) We see differently:  the pigeonhole has a "detection" in any EUVE bandpass
	other than Lexan but the source was detected in the WFC S1 bandpass.

Fig. 2 shows graphically the categorization algorithm for comparing EUVE with
WFC sources.
   The procedure detect finds a "detection" in a filter when the attribute de-
tection_quality from the fil table is greater than a predefined threshold value.
This threshold value is given as a parameter to the stored procedure when it
is invoked.  The detections in each filter are also visually verified and the
results are entered into the database to verify the results from detect.


4. SUMMARY

   EUVECATDB has a general purpose schema design and can be used for multiple
purpose queries.  With the ADS interface, both QBE (Query By Example) and SQL
queries can be used.  While the survey data will remain the property of Berkeley
for a period of one year following survey gap-filling and will not be available
to the public, the Bright Source List that became public in June 1993 will be
accessed through the EUVE Archive [4].


REFERENCES

1. H. L. Marshall, in Data Analysis in Astronomy III, ed. V. di Gesu et al.,
	New York, Plenum, pp. 169-177 (1989).
2. R. F. Malina, et al., The Extreme Ultraviolet Explorer Bright Source List,
	Publication of the Center for EUV Astrophysics, University of Califor-
	nia at Berkeley (1993).
3. B. Antia, JBIS, this issue (1993).
4. J.J. Drake, C. Dobson, E. Polomski, "The Extreme Ultraviolet Explorer Ar-
	chive", to appear in Astron. Soc. Pac. Conf. Series (1993).


FIGURE CAPTIONS

Figure 1:  EUVECAT Database Schema
Figure 2:  Algorithm "detect"