#shell and #sql

The #sql script directive invokes an SQL command and captures the results. The #sql directive is currently available only for builds related to the QUISP/SHSQL package. Interfaces to other databases (MySQL, Oracle, etc.) not yet available but should be possible by coding in dbinterface.c.

Note that both of these are of limited value in ploticus... proc getdata already has a command attribute for getting data from shell command results, and it also has a sql attribute for running an SQL command. They should be viewed as providing an alternative approach.

STREAMING DATA

	#proc getdata
	delim: tab
      data:
	#sql #dumptab select * from trialresults where trialnum = @TRIALNUM

For longer SQL commands a multi-line construct is available:

	#proc getdata
	delim: tab
      data:
	#sql #dumptab 
		select trialnum, age, sex, weight, time1, time2
		 from trialresults 
		where trialnum = @TRIALNUM and trialdate = @TDATE
      #endsql

SETTING VARIABLES

	#sql #load select fullname from tsyms where tickersym = @TSYM

      #proc annotate
      location: min max-0.2
      text: @fullname

Here is an example where the results of a shell command may be used to set variables:

      // submit the command..
      #shell #processrows cat mydata | grep "^Jan"
      // get one result rows and parse out fields..
      #call $shellrow( A, B, C ) 

	#proc annotate
	location: min max-0.2
	text: @A @B

DRIVING LOOPS

	// get a list of cases
	#sql #processrows select caseid from cases
	#while $sqlrow() = 0
	  // get data for a case
	  #proc getdata
	  delim: tab
	  data:
	    #sql2 #dumptab select * from casedata where caseid = @caseid
	    ... plot the case

	#endloop

#shell may be used similary.