after xplore installation java code is not working for full text search. it brings back zero result

Hi,

After we migrated to xplore from Fast we are experiencing full text search troubles. The piece of java code for full text search which was working perfectly alright with Fast is now bringing back zero result. So then I tried a sample full text search code from documentum developer's guide but it is also not working. When we tried the DQL it uses on the webtop then it seems to fetch results. There is no exception and error in the log. The sample code is something like this: -

publicString[] searchDocument(String searchText) throws CusotmException{ String[] docIds =null; List documentObjectIds =new ArrayList();

DomainManager domainManager = DomainManager.getInstance();

String docbaseUserId = domainManager.getDocBaseUserId();

String docbasePassword = domainManager.getDocBasePassword();

this.session = obtainSession(docbaseUserId, docbasePassword);

StringBuffer query =new StringBuffer(); query.append("select r_object_id from dm_document "); query.append("SEARCH DOCUMENT contains '"); query.append(searchText);

query.append("' WHERE (a_is_hidden = FALSE)"); System.out.println(query.toString()); IDfClientX clientx =new DfClientX(); IDfQuery queryObject = clientx.getQuery();

queryObject.setDQL(query.toString());

IDfCollection coll;

IDfId idObj =null;

try {coll = queryObject.execute(

this.session, IDfQuery.DF_READ_QUERY );

int i = 0;

while (coll.next()){

idObj = coll.getId("r_object_id"); documentObjectIds.add(idObj.getId());

System.out.println(i + " ID: " + idObj.getId()); i++;

}

catch (DfException dfe) {

try {

this.session.disconnect();}

catch (DfException dfe1) {

throw new CustomException(dfe1, "Failed to disconnect session");}

throw new CustomException(dfe, "Failed to get objects for the text criteria");}

try {

this.session.disconnect();}

catch (DfException dfe1) {

throw new CustomException(dfe1, "Failed to end session");}

docIds =

new String[documentObjectIds.size()];

return (String[]) documentObjectIds.toArray(docIds);

}

Does anybody know anything about this?

Find more posts tagged with

Note:

If there are Accepted Answers, those will be shown by default. You can switch to 'All Replies' by selecting the tab below.

Accepted answers

DCTM_Guru

Assuming that you are using the same username/password as in DA, are you sure that your custom code is establishing a valid session?

All comments

DCTM_Guru

Does your DQL return anything if you run it in DA? If not, then its not your java code, which means that the full text index might need to be rebuilt.

HogWild

Yes, it returns results.

Also wanted to tell you what we are using currently: -

- Documentum xPlore (Search Services) Version: 1.0

- Content Server version: 6.6 (Migrated from 5.3 SP6 replacing FAST indexing)

- Operating System version of CS: Windows Server 2008 R2 Standard 64 Bit 8GB RAM

I am not sure if any of the above combination is causing any trouble.

DCTM_Guru

Assuming that you are using the same username/password as in DA, are you sure that your custom code is establishing a valid session?

HogWild

Yes, I think so. I am creating a IDfSession here. I am sure a session is created successfully because when I tried a DQL something like "Select * from dm_document" then it happens to work. Note this is a very simple DQL and does not contain full text. So this makes me believe a session is getting created successfully.

Is there a way to verify this is not the right session I should be using?

DCTM_Guru

Whats the actual DQL that is being executed (ie what value is getting passed as searchtext)? Can you system.out DQL statement?

HogWild

the DQL is:

select r_object_id from dm_document SEARCH DOCUMENT contains 'document' WHERE (a_is_hidden = FALSE)

We also tried for different texts like 'rat', 'EMC', etc but it comes back with zero result. These are texts which should be found in the documents we have in our repo.

Do you think I should try some other Java approach to do the full text search? I am not sure which one. Please advise.

DCTM_Guru

What happens if you drop the WHERE clause?

HogWild

It did not make any difference. I think I should get the full text indexes rebuilt. I am not sure but I think I should try this step.

mszurap

Have you tried to issue the search on the xplore administration console? You should double check whether IndexAgent and xplore Index Server is working well.

HogWild

Hi, yes I tried issuing search from xplore admin and it works there. Searches are even working from da. I also verified IndexAgent and xplore index server are running good.

For instance, today I added a new PDF in the docbase through da and it was indexed instantly and then I searched one unique word from that PDF at both the locations i.e. xplore admin and da and it searched me the PDF. Then I tried the same text from my java application but it came back with zero result. No exception in the log too.

Not sure what should I try next.

DCTM_Guru

Does your DFC app returns rows if you just do normal search:

select * from dm_sysobject where r_creation_date>date(yesterday())

HogWild

I tried this from da: select * from dm_sysobject where r_creation_date>'2011-05-09' and it fetched some rows. But i think these rows are not the real documents or maybe nothing was added to the docbase after '2011-05-09'. I have pasted one row below for your reference. First column is object id, second is object name, third is title and fourth is subject.

09009210800c081e

5/9/2011 12:10:49 AM dm_DataDictionaryPublisher

Result of dm_method(dm_DataDictionaryPublisher) with status code (0)

Result of dm_method(dm_DataDictionaryPublisher) with command line: .\dmbasic -f..\install\admin\mthd7.ebs -eDataDictionaryPublisher -h 7296 -- -docbase_name TOXDOCS.TOXDOCS -user_name USF....

It seems the query works in da like the other dqls i tried worked too in da.

HogWild

Thanks for helping out. We figured out the user which we were using was have zero privillege and then we added superuser privillege to it and it worked then.

mszurap

One more thing which can help one in case of doubt. I found it in xPlore 1.0 Deployment Guide, Page 61:

Verify that search executed against xPlore

To check whether a search is executed against FAST or xPlore, use the ENABLE(FTDQL_DATA) hint.

For example, issue this iAPI command (one a single line):

?,c,SELECT r_object_id FROM dm_sysobject SEARCH DOCUMENT CONTAINS ’dmadin’ ENABLE(FTDQL_DATA)

The result will contain an XQuery snippet if the Content Server executes against xPlore. For example:

...Native Query:[((. ftcontains ( ((’dmadmin’) with stemming) ) ))and ( ( (dmftinternal/i_all_types = ’0300271080000105’) )and ( (dmftversions/iscurrent = ’true’) ) )]