topic Re: Lucene and indexing in Alfresco Archive

Lucene and indexing

bgl — Fri, 04 Feb 2011 00:04:38 GMT

Hello,I'm using Alfresco 3.4d community edition.I have created a folder (using FileFolderService API) in the company home. When I search with the NodeBrowser, I have very strange results :+PATH:"/app:company_home/*"I get the list of folders as result. This is completely normal. Now, I try this :+PAT

Re: Lucene and indexing

bgl — Fri, 04 Feb 2011 00:06:18 GMT

Note that the following query is returning the correct result :

+PATH:"/app:company_home/cm:folderName"

So why is this running fine and not the previous ones ?

Thank you.

Re: Lucene and indexing

bgl — Fri, 04 Feb 2011 14:09:55 GMT

My folder name has an underscore which seems to be a problem with standard configuration ( see this : http://forums.alfresco.com/en/viewtopic.php?f=27&t=21955 )

So I think my problem is almost solved 🙂
I will try the Luke tool to read into the lucene indexes.

Boris

Re: Lucene and indexing

openpj — Fri, 04 Feb 2011 14:20:01 GMT

Remember that when you are searching using the PATH token with Lucene you have to use the ISO9075 encoding for all the space names:
If you are using the JAVA API for example you have to encode the space name in this way:


…
String folderName = "anyName_for_yourSpace using also blankspace characters";
String luceneQuery = "PATH:\"/app:company_home/cm:\""+ISO9075.encode(folderName)+"\"";
…
‍‍‍‍‍‍

Alfresco stores nodes using XPATH and that's why you have to conform to this standard and you have to encode all the space names using the ISO975 format.

Re: Lucene and indexing

bgl — Fri, 04 Feb 2011 14:50:14 GMT

Hi,

thank you for your answer. Yes that's true I had completely forgotten this thing in ISO9075. I had used it before but did not remember this "detail".
I will try it and post the result.

Again thank you !

Re: Lucene and indexing

bgl — Wed, 09 Feb 2011 12:59:23 GMT

Hi,

ISO is not the problem (well not in my case). The problem is about tokenization of file name… I tried a config with stored = true in the content model but it does not seem to change something…

I am still investigating the different configuration parameters to get something that is ok for me…

Boris

Re: Lucene and indexing

mrogers — Wed, 09 Feb 2011 13:38:00 GMT

If you are using FileFolderService I suggest you use searchSimple instead. Its far faster.