onpladm load job with spec file (includes simple but lengthy test case)

Part of the fun of migrating to new releases (besides learning new bugs and incompatibilities with old apps) is learning the new utilities. For example, I am attempting to learn the onpladm tool which provides a command-line interface to the High Performance Loader (HPL) with IDS 9.40.UC2 on Solaris. I am attempting to create a project and a load job through spec files. So far, failure occurs when reading the LOADJOB portion of the spec file. The scripts I started with were provided by someone who claims they work on IDS 9.20.TC1, a version unavailable to me. In the end, though, I built my own scripts using the highly uninformative IBM Informix High-Performance Loader User's Guide, Version 9.4 (G251-1255-00) available online at: http://publibfi.boulder.ibm.com/epubs/pdf/ct1t3na.pdf It still does not work. So far, support has not been able to answer my question, which is fairly simple: am I doing something wrong, or is this a bug? Of course, once I get the answer to that, I hope to find out how to fix the problem. Unfortunately, I am not just doing this for fun; I have a deadline by which I have to have an answer. As if by coincidence, it is Wednesday. If anyone would care to peruse my case and see if they can provide an answer, I would be most appreciative. I have included my test case after my signature (because it is too long and no one would read that far down, if it were at the top). Sincerely, Christopher Coleman Database Analyst Pharmacy Division Mediware Information Systems, Inc. Here are the commands I run; beware of the first one: echo "DROP DATABASE onpload ;" |dbaccess -e echo "DROP DATABASE targetdatabasename ;" |dbaccess -e echo "CREATE DATABASE targetdatabasename WITH LOG;" |dbaccess -e echo "CREATE TABLE targettablename (columnname CHAR);" |dbaccess -e targetdatabasename onpladm create project projectname onpladm create object -F specfilename Here are the specfilename contents. BEGIN OBJECT DEVICEARRAY device_array_name BEGIN SEQUENCE TYPE FILE FILE device_path TAPEBLOCKSIZE 0 TAPEDEVICESIZE 0 PIPECOMMAND END SEQUENCE END OBJECT BEGIN OBJECT DELIMITEDFORMAT formatname PROJECT projectname CHARACTERSET RECORDSTART RECORDEND FIELDSTART FIELDEND FIELDSEPARATOR | BEGIN SEQUENCE FIELDNAME fieldname FIELDTYPE Chars END SEQUENCE END OBJECT BEGIN OBJECT LOADMAP mapname PROJECT projectname FORMAT formatname DATABASE targetdatabasename TABLE targettablename BEGIN SEQUENCE COLUMNNAME columnname FIELDNAME fieldname JUSTIFICATION CASECONVERT DEFAULTVALUE TRANSFERBYTES COLUMNOFFSET FIELDOFFSET FIELDMINIMUM FIELDMAXIMUM FILLCHARACTER PICTURE FUNCTION STORAGECODING BLOBCOLUMN END SEQUENCE END SEQUENCE END OBJECT BEGIN OBJECT LOADJOB jobname PROJECT projectname DEVICE device_array_name MAP mapname FILTER SERVER DATABASE REJECTFILE LOGFILE RUNMODE GENERATEVIORECS TAPES NUMRECORDS STARTRECORD MAXERRORS END OBJECT

On 24 Nov 2003 10:06:50 -0800, christopher.coleman@mediware.com (Christopher) wrote: >Part of the fun of migrating to new releases (besides learning new >bugs and incompatibilities with old apps) is learning the new >utilities. > >For example, I am attempting to learn the onpladm tool which provides >a command-line interface to the High Performance Loader (HPL) with IDS >9.40.UC2 on Solaris. > >I am attempting to create a project and a load job through spec files. > So far, failure occurs when reading the LOADJOB portion of the spec >file. > >The scripts I started with were provided by someone who claims they >work on IDS 9.20.TC1, a version unavailable to me. In the end, >though, I built my own scripts using the highly uninformative IBM >Informix High-Performance Loader User's Guide, Version 9.4 >(G251-1255-00) available online at: >http://publibfi.boulder.ibm.com/epubs/pdf/ct1t3na.pdf Something that may help would be the (hopefully more informative) article written about onpladm (from yours truly) at the Informix Developer Zone . . . http://www-106.ibm.com/developerworks/db2/zones/informix/library/techarticle/0307carlson/0307carlson.html

John Carlson offered his more informative article about onpladm: > http://www-106.ibm.com/developerworks/db2/zones/informix/library/techarticle/0307carlson/0307carlson.html Actually, John, I had already read, bookmarked, and printed out that article, which was very helpful. Unfortunately, it does not give an example of creating load jobs with spec files, just device arrays. It is the load job which fails with a syntax error, despite my best efforts. John uses the term "template file", the manual uses the terms "Specification-File" or "spec file". I, personally, have been using a whole different class of terms. Sincerely, Christopher Coleman Database Analyst Pharmacy Division Mediware Information Systems, Inc.

Christopher wrote: > > Part of the fun of migrating to new releases (besides learning new > bugs and incompatibilities with old apps) is learning the new > utilities. > > For example, I am attempting to learn the onpladm tool which provides > a command-line interface to the High Performance Loader (HPL) with IDS > 9.40.UC2 on Solaris. > > I am attempting to create a project and a load job through spec files. > So far, failure occurs when reading the LOADJOB portion of the spec > file. > > The scripts I started with were provided by someone who claims they > work on IDS 9.20.TC1, a version unavailable to me. In the end, > though, I built my own scripts using the highly uninformative IBM > Informix High-Performance Loader User's Guide, Version 9.4 > (G251-1255-00) available online at: > http://publibfi.boulder.ibm.com/epubs/pdf/ct1t3na.pdf > > It still does not work. > > So far, support has not been able to answer my question, which is > fairly simple: am I doing something wrong, or is this a bug? Of > course, once I get the answer to that, I hope to find out how to fix > the problem. > > Unfortunately, I am not just doing this for fun; I have a deadline by > which I have to have an answer. As if by coincidence, it is > Wednesday. > > If anyone would care to peruse my case and see if they can provide an > answer, I would be most appreciative. I have included my test case > after my signature (because it is too long and no one would read that > far down, if it were at the top). > > Sincerely, > > Christopher Coleman > > Database Analyst > Pharmacy Division > Mediware Information Systems, Inc. > > Here are the commands I run; beware of the first one: > echo "DROP DATABASE onpload ;" |dbaccess -e > echo "DROP DATABASE targetdatabasename ;" |dbaccess -e > echo "CREATE DATABASE targetdatabasename WITH LOG;" |dbaccess -e > echo "CREATE TABLE targettablename (columnname CHAR);" |dbaccess -e > targetdatabasename > > onpladm create project projectname > onpladm create object -F specfilename > > Here are the specfilename contents. > [ ... snip ... ] > BEGIN OBJECT LOADJOB jobname > > PROJECT projectname > DEVICE device_array_name > MAP mapname > > FILTER > SERVER > DATABASE > REJECTFILE > LOGFILE > RUNMODE > GENERATEVIORECS > TAPES > NUMRECORDS > STARTRECORD > MAXERRORS > > END OBJECT what ist the error? At the point of error do you have *anything* in your $DBONPLOAD database? Does onpladm list device give you some output? If yes, does onpladm decribe device device_array_name give you output, which looks as expected? Also try to list all other entries (formats, maps ...) Here is, what I do: I use the quick, command based way to get sort of 'skeleton' (describe puts it out in nice form) I never did combine more than 1 step: I have separate onpladm calls for device first, then for generate the jobs, then for running the jobs. Only in the case of PIPECOMMAND one must edit the contents of the dsf file, as the first whitespace ends the text (a well known behaviour) Else I never saw any problem. dic_k -- Richard Kofler SOLID STATE EDV Dienstleistungen GmbH Vienna/Austria/Europe