Checking the loading process ============================ Procedure for one CD -------------------- 1. Get last patent of exchange.idx of CDROM using tail command 2. Check if patentnumber is present in the MAIN table of the database using the Patent Number Search screen 3. Check if patentnumber is present in Verity using the Basic Search screen 4. Check if patentnumber is present in imageserver idx file by viewing the image in the detail view Procedure for checking the whole system --------------------------------------- All these commands should be executed as user ipsadmin. Image server: ------------- To get a list of all CD's in the jukeboxes and store the result in file /arc/convdata/cdjukes.lst, enter 'ls' -1 /cdrom > /arc/convdata/cdjukes.lst To get a list of all CD's, for which an image server index has been created and store the result in file /arc/convdata/cdimgsrv.lst, enter grep DISC /ips/idx/jukebox.parm | cut -d= -f2 | cut -d, -f1 | sort > /arc/convdata/cdimgsrv.lst Using these two files, a list of CD's which are in the jukeboxes but not in the imageserver indexes or which are in the imageserver indexes but not in the jukeboxes, can be obtained by entering diff /arc/convdata/cdjukes.lst /arc/convdata/cdimgsrv.lst All lines starting with "<" are CD's which are in the jukeboxes but not in the imageserver indexes. Lines starting with ">" are CD's which have been removed from the jukeboxes but not from the imageserver indexes. Webserver: ---------- To get a list of all CD labels in the database and store the result in the file /arc/convdata/cddb2.lst, enter: db2 "connect to patent" db2 "select distinct text_cd from inst1.main order by text_cd" | tail +4 | cut -f 1 -d " " | grep -v "^$" > /arc/convdata/cddb2.lst Using this file and the file cdimgsrv.lst created above, a list of CD's which are in DB2 but not in the imageserver indexes or which are in the imageserver indexes but not in DB2, can be obtained by entering diff /arc/convdata/cdimgsrv.lst /arc/convdata/cddb2.lst All lines starting with "<" are CD's which are in the imageserver indexes but not in DB2. Lines starting with ">" are CD's which are in DB2 but not in the imageserver indexes. To get one patentnumber of each CD in the database and store the result in the file /arc/convdata/patndb2.lst enter: db2 "connect to patent" db2 "select min(patn) from inst1.main group by text_cd" | tail +4 | cut -f 1 -d " " | grep -v "^$" > /arc/convdata/patndb2.lst To check if these patents are indexed correctly in Verity, enter mkvinspect /arc/convdata/patndb2.lst This creates the executable script dovinspect. Now enter dovinspect This searches in Verity for each patent listed in the input file. The output of the script is stored in the file db2vpatn.out. If a patent is listed more than once in this output file, it is indexed more than once in Verity and one should run bibdelidx followed by bibupdidx for the CD for this patent which is listed in the MAIN table. If a patent is missing in this output file, it is not indexed in Verity and one should run bibupdidx for the CD. Now enter uniq db2vpatn.out > /arc/convdata/uniq.lst To store the missing patents in file /arc/convdata/patnvnot.lst, enter diff /arc/convdata/uniq.lst db2vpatn.out | grep ">" | cut -f 2 -d " " > /arc/convdata/patnvnot.lst To store in /arc/convdata/patnvnotdb.lst a list of patents which are in Verity, but not in DB2, enter diff /arc/convdata/uniq.lst db2vpatn.out | grep "<" | cut -f 2 -d " " > /arc/convdata/patnvnotdb.lst To store the double patents in file /arc/convdata/patnvdouble.lst, enter diff /arc/convdata/uniq.lst db2vpatn.out | grep ">" | cut -f 2 -d " " > /arc/convdata/patnvdouble.lst The above three files contain patent numbers, not CD labels. Use the command whichlabel.pl to get a list of CD labels for each patent listed in the input file (this scripts uses the MAIN table). Note: scripts can be found on spider:/home/ipsadmin March 21, 2000 Sander Berents