RE: SQL from korn shell

In article <71b0fd$svm$1@news.xmission.com>, rbernste@alarismed.com (Bernstein, Rick) wrote: > > Being a novice with korn shell script programming, I pipe the results > of SQL command to "tail" and "head" commands to eliminate extra > blank lines. For example: > > > integer SYSUTILS_ROWNUM=`dbaccess sysutils 2>/dev/null <<- EOF! > OUTPUT TO PIPE 'tail +3 | head -1' WITHOUT HEADINGS > SELECT count(*) > FROM bar_instance > WHERE ins_aid < $1;> EOF!` > > It successfully returns a numeric value to the shell script variable. > However, someone else may have a better technique. > > > -----Original Message----- > From: mr_mark@my-dejanews.com [mailto:mr_mark@my-dejanews.com] > Sent: Thursday, October 29, 1998 13:40 > To: informix-list@iiug.org > Subject: SQL from korn shell > > > I am trying to get a value from a database. I couldn't get the SELECT > call > to just return a data value without the preceding newlines so I have the > OUTPUT TO command sending the output to a file which I then read to > eliminate > the newlines. All this is initiated from an awk script which has stdout > redirected to a file. The SELECT is in a korn shell script. The > problem > is: > even if the SQL table query returns no value I get a newline printed to > stdout(my file) and it's not being done by me. Somehow, a dbaccess SQL > call > automatically behind the scene puts a newline to stdout. Is there a > way to > supress this newline and better yet to get the original value from the > SQL > call into a variable. My code is below: # Arguments coming include > database, > field name being retreived, table to query # field to match key to and > the > key gotten from the log file. These arguments # must come in this > order. > > dbaccess $1 - - << EOF 2>/dev/null > OUTPUT TO sqlout WITHOUT HEADINGS > SELECT $2 > FROM $3 > WHERE $4 = $5; > EOF > > while read -r line > do > if [[ -z $line ]] > then > line="" > continue > fi > > print -n "$line" > testfile > break > done < sqlout > rm -f sqlout > > -----------== Posted via Deja News, The Discussion Network ==---------- > http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own > > Spooky something was asked about this yesterday try NUMFND=`dbaccess ${DBNAME} -<<EOF 2>/dev/null | awk '/ [0-9]/{print $1}' select count(*) from systables where tabname = "{$TABNAME}"EOF ` It only works with numeric obviously. If you selecting column data then a '^$' filter in sed or egrep will be a lot quicker than head and tail Paul Watson # I don't suffer from WF Software Ltd. # stress, I'm just Tel. (+44) 1436 674729 # a carrier Fax. (+44) 1436 678693 # www.wfsoftware.co.uk