Making ODBC easier By Dave Doulton University of Southampton
Jan 11, 2016
Making ODBC easier
By
Dave Doulton
University of Southampton
Agenda
• Introduction
• Setting up ODBC sources
• ODBC import
• ODBC pql
• ODBC members
• Conclusion
Introduction
• ODBC Open DataBase Connectivity
• This session will show you how to set up ODBC sources and use them from SIR.
• It will show the simplest way to go about this with PQL members available from SUG website.
Setting up ODBC sources
• ODBC sources are set up using Data Sources from the Administrative Tools Menu
• Sources can be User, available to that user only or System, available to all users both are configured in exactly the same way but from different tabs
Setting up ODBC sources
• I shall show the steps in setting up some System DSNs
• I have previously set them all up as User DSNs for use in the demonstrations
• Click add on the screen shown prior to this
Setting up ODBC sources
• Choose the source and click Finish.
• You will then get a configuration screen which will vary according to the source.
• The next screens will show various different types
SIR
SIR different server
CSV
CSV options
CSV Define Format
Excel with options
Access with options
Access advanced options
FoxPro with options
dBase with options
Paradox with options
Oracle
All defined
ODBC import
• Data can be imported into a SIR database from an ODBC source very easily especially into a caseless database as there is no Case Id to worry about.
• Create a caseless database in this case TESTODBC and choose import records from the database menu
Choose ODBC and choose data source
Select Source
This gives
Clicking Next gives:- note sheet name with a $ and a range
Clicking Next gives
Change record name and choose roword as sortid. This gives each row a new number
Clicking OK creates record and imports.
ODBC import
• This is fine if you want all the data in a database.
• But I find that I want to read a data file select some data from that file and then do something with it without storing it in another database. E.g. for running procedures on an Oracle database.
• In these cases PQL is the answer.
ODBC pql
• The ODBC import is based on a set of pql statements.
• These seem a bit daunting at first.
• See the next screens for details.
Connect and Disconnect
• CONNECT conid SERVER name {DATABASE name | TABFILE name} [USER name] [PASSWORD name] [PREFIX name] [UPDATE | READ] [ERROR errid]
• DISCONNECT conid [ERROR name]
Statement, Prepare Bind Execute and Delete• STATEMENT statid CONNECT conid [ERROR name]
• PREPARE STATEMENT statid CONNECT conid {COMMAND text_expression | BUFFER buffer_name} [ERROR name]
• BIND STATEMENT statid CONNECT conid (param_no,value)
• EXECUTE STATEMENT statid CONNECT conid [ERROR name]
• DELETE STATEMENT statid CONNECT conid [ERROR name]
An Example
See the bottom of the help on ODBC at
file:///C:/Program%20Files/SIR2002/help/visualpql_odbc.htm
Something easier would clearly be nice.
ODBC members
• I have created a simple set of members that can be put in the sysproc ODBC family and make life a lot easier.
• It includes a simple template member as shown on the next screen
Template
• program• call sysproc.odbc.header• compute srce=''• compute user=''• compute pwd=''• call sysproc.odbc.connect• compute qtext=''• call sysproc.odbc.loop ($C $)• write cname(1) 5e• write fields(1) 5e• call sysproc.odbc.endloop ($C $)• call sysproc.odbc.disconn• call sysproc.odbc.tail• end program
The template filled in
• program• call sysproc.odbc.header• compute srce='1 CSV'• compute user='dcd'• compute pwd='dcd'• call sysproc.odbc.connect• compute qtext='select * from employee.csv'• call sysproc.odbc.loop ($C $)• write cname(1) 5e• write fields(1) 5e• call sysproc.odbc.endloop ($C $)• call sysproc.odbc.disconn• call sysproc.odbc.tail• end program
A more complicated example
• program• DEFINE PROCEDURE VARIABLES ARRAYS• string ofns• call sysproc.odbc.header• execute subroutine sysproc.menu.odbclist returning(srce)
• display textbox "Enter output filename" response ok,ofns• if(ok <=0) exit program• if(srst(ofns,".csv")=0)compute ofns=ofns+'.csv'
• display textbox "Enter username" response ok,user• if(ok <=0) exit program• display textbox "Enter password" secret response ok,pwd• if(ok <=0) exit program
• call sysproc.ODBC.CONNECT• display textbox "Enter query" response ok,qtext• if(ok <=0) exit program
• call sysproc.odbc.loop | ($C $)• perform procs• call sysproc.odbc.endloop• call sysproc.odbc.disconn• call sysproc.odbc.tail• report filename=consol/sort=rnum
before reportstring*4000 outlineopen ofn/dsnvar=ofns/write/lrecl=4000compute outline=""for k=1,cnumcompute outline=outline+'"'+trim(cname(k))+'"'if(k<cnum)compute outline=outline+','end forwrite(ofn)outlinebreak level 1, rnumdetail blockcompute outline=""for k=1,cnumifthen(exists(fields(k)))compute outline=outline+'"'+trim(fields(k))+'"'elsecompute outline=outline+'" "'endifif(k <cnum)compute outline=outline+','end forwrite(ofn) outlineend break levelafter reportclose ofnpql escape ['"c:\windows\system32\cmd" /c "start '+ofns+'" '] wait 0 minimize 1end reportend program
An update example
• program• call sysproc.odbc.header• compute srce=''• compute user=''• compute pwd='‘• Compute utext=‘’• call sysproc.odbc.connect• call sysproc..odbc.update• call sysproc.odbc.disconn• call sysproc.odbc.tail• end program
Update filled in
• program• call sysproc.odbc.header• compute srce=‘3 Access'• compute user=‘dcd'• compute pwd=‘dcd‘• Compute utext=‘update employee set ndepends=4’• call sysproc.odbc.connect• call sysproc..odbc.update• call sysproc.odbc.disconn• call sysproc.odbc.tail• end program
Web access
• With slight alterations these can be used via the web as well.
• See the following example.
programDEFINE PROCEDURE VARIABLES ARRAYSstring ofnscall sysproc.odbc.headercompute srce=cgivarpn("srce")compute user=cgivarpn("user")compute pwd=cgivarpn("pwd")call sysproc.ODBC.CONNECTcompute qtext=cgivarpn("qtext")call sysproc.odbc.loop ($C $)perform procscall sysproc.odbc.endloop ($C $)call sysproc.odbc.disconncall sysproc.odbc.tailreport filename=consol/sort=rnum
before reportstring*4000 outlinewrite(cgi)['<h2 >Output from '+srce+'</h2>']write(cgi)['<h3 >Using query '+qtext+'</h3>']compute outline="<table border=><tr >"for k=1,cnumcompute outline=outline+'<th >'+trim(cname(k))+'</th>'end forcompute outline=outline+'</tr>'write(cgi)outlinebreak level 1, rnumdetail blockcompute outline="<tr >"for k=1,cnumifthen(exists(fields(k)))compute outline=outline+'<td >'+trim(fields(k))+'</td>'elsecompute outline=outline+'<td > </td>'endifend forcompute outline=outline+'</tr>'write(cgi) outlineat end blockend break levelafter reportwrite(cgi)'</table>'end reportend program
Notes
• When using sources that can have different file extensions then use the extension in the tablename for select.
• In Excel use the sheet name with a $ or a named range
• In SIR the password can be all 4 passwords separated by commas
• i.e. userpass,company,high,high userpass can be anything but is needed always
Conclusion
• I hope you will all have a try of these routines to speed up your access to other data.
• I know it has speeded up my development work no end.