arsload need generic indexer?

[kbsiva]

Hello,
  I am trying to do an arsload from AIX6.1 machine, to load a standard oracle ascii report. The command I am using is,

/usr/lpp/ars/bin/arsload -u aaa-p aaa -h CON -b request_id -B AG.APP.IGN.IDX.EXT  CMSTEST.CMSTESTAPP.120214.124050.txt

and the error is,

arsload: Unable to add another index field to the data.  Data may be fully-composed or otherwise already indexed.
arsload: Processing failed for file >CMSTEST.CMSTESTAPP.120214.12405042.txt<
arsload: Processing has stopped.  The remaining files will NOT be processed.

My questions are as follows: Do you need a generic index parameter file defined to load ascii data? I was hoping some how to tell arsload to use the indexer(IGN) on the file name instead of reading from the report itself.
 Is there a way we can tell arsload to create an index based on the file name instead of using the indexer?

Thanks and appreciate your help in advance,

Siva

[kbsiva]

And yes we are in 8.4.1

[jeffs42885]

This is what I do, hope this provides some insight...

/usr/lpp/ars/bin/arsload -h host -u user -p password -nvf -X G -g appGroup -a Appname App.0.6401D.86401D

Try removing the .txt from the index file see if that helps, or maybe try the above sample that I use.

[kbsiva]

Hi Jeff,
 Thanks for the response.
 I tried that and it still the same error. We created the App Group and App from the Admin Client utility running in Windows. Do you know what data type to specify for Ascii data? I choose User defined and also I tried by setting the field "request_id" to be the indexer in the App Group. Same result. I then tried to set the field request_id to be the filter and same result. So is there anything specific we need to define in the App Group/App ? Is there a unix command that we can use to set up App Group/App ?

[kbsiva]

Hi,
 I am getting the error with -X G

arsload: Indexing started , 14765 bytes to process
arsload: Unknown Indexer ><
arsload:Output Indexer was not created
arsload:Indexing failed
...

[Alessandro Perucchi]

Hello Siva,

One question, how did you configure your application "Appname" ??

ACIF Indexer?
Generic Index?

Something else?

If you have put it to "Generic Index", then you MUST provide an index file with your txt file.
If you have defined "ACIF Index", then it should work without problems. But since it doesn't work, then I suspect you haven't define it.

Sincerely yours,
Alessandro

[kbsiva]

Thanks for the reply Alex
 I am getting it to load using the ACIF indexer graphical tool and the Generic Indexer parameter file. My problem is this,

arsload -u name -p passwd -h server -g APPG -a APP -b "workload" -B AG.APP.IDX.IGN.EXT
APPG.APP.A123.C120202.TXT

In my application group , I have two indexes created workload and reqid . One index(reqid) I want to use for the the data inside the file. The other index I want to use from the file name. I have already created the trigger, field, index(reqid) for the data inside the report. That index is getting created fine when i do the load. I am just trying to get the arsload to take the second index value from the -b parameter in the command line.

[Alessandro Perucchi]

Hello Siva,

Well now I've understood what you are trying to do :-D before it was a bit not really clear (or maybe I've not read correctly...) sorry.

Well, I think you are going to be desapointed...

Here is what the arsload documentation says: (http://publib.boulder.ibm.com/infocenter/cmod/v8r5m0/topic/com.ibm.ondemand.administeringmp.doc/dodlo003.htm)

The IDX value works with the IBM Content Manager OnDemand PDF Indexer for Multiplatforms only. It does not work with the OS/400 indexer or the generic indexer.

so if you are trying to use the option -b and -B with ACIF indexer, you're out of luck. Since that option works ONLY with PDF Indexer.

In addition to that, in the same documentation page as before:

You can use the -B parameter or the -g and -a parameters combined but cannot use all three of them together. If you specify the -B parameter, you identify the application group name and the application name as parts of the filename.

Meaning that either you use the -B option, or (exclusive or) you use -g & -a. But not at the same time.

So in your case, you want to use that format of arsload:

arsload -u name -p passwd -h server -b "workload" -B AG.APP.IDX.IGN.EXT APPG.APP.A123.C120202.TXT


What you could do (I've done it a couple of times), is to write a small script and extract your IDX from the filename, and write it in a file (eg: index-param.prm) with the following format:

FIELD32=A123     <- value of the IDX from the filename
INDEX32=workload,FIELD32

and then call the arsload command like that:

arsload -u name -p password -h server -g APPG -a APP -j index-param.prm APPG.APP.A123.C120202.TXT

Just take care to use an INDEX/FIELD number that is high enough not to have the same number as your actual Acif Indexer.

I hope that would help a bit.

Sincerely yours,
Alessandro

[kbsiva]

Ok this what I did and this works for me. This is in AIX6.1, CMOD8.4. The custom oracle reports had some control characters, like ^L and ^M. I removed them first, using the tr -d '\f' <filename> file and , tr -d '\n' <file> file.
Next I used the acif indexer to define two indexes (reqid, workload) using sample data.
Next I used the arsload like this

arsload -u admin -p passwd -h CON -g AG -a APP -X G -nvf -b workload -B AG.APP.IDX.IGN.IGN.EXT AG.APP.0102.121202.302.TXT

so what this is doing is basically over riding the workload value in the file with the one I am supplying from the file name. So the index is now 0102 from the instead of the value from inside the file.