Mailing-List: contact lucene-user-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Lucene Users List" <lucene-user@jakarta.apache.org>
Received-SPF: neutral (hermes.apache.org: local policy)
Message-ID: <418F43F1.1080102@intrafind.de>
Date: Mon, 08 Nov 2004 11:01:21 +0100
From: Bernhard Messer <bernhard.messer@intrafind.de>
User-Agent: Mozilla Thunderbird 0.8 (X11/20040913)
MIME-Version: 1.0
To: Lucene Users List <lucene-user@jakarta.apache.org>
Subject: Re: Transaction in Lucene
References: <20041107140941.32709.qmail@web12704.mail.yahoo.com>
In-Reply-To: <20041107140941.32709.qmail@web12704.mail.yahoo.com>
Content-Type: multipart/alternative;
 boundary="------------030902080302050800010400"

--------------030902080302050800010400
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

The message "No tvx file" can appear, if you have term vectors enabled during index and the documents you are adding have empty fields. As an example, if you try to index html documents, where many of them don't have a valid html title, the message will raise up. Looking at the term vector relevant code, this is nothing you have to worry about, it is just a status message. Otis is right, it is planned for future releases to avoid System.out.println() statements within lucene.

regards
Bernhard 


Otis Gospodnetic schrieb:

>I'm not sure about the tvx error, but I think I recall somebody
>changing some code around it a month or two ago.  I also believe
>System.out.println is on the TODO list for elimination.
>
>Otis
>
>--- commandor <commandor@inbox.ru> wrote:
>
>  
>
>>Hello,
>>
>>I came across the following problem with "No tvx file". 
>>
>>How could I manage to get it?
>>I like to have transaction processes in Lucene.
>>After my reading dev-lucene and user-lucene lists and analysing what
>>people suggested
>>I made up my own.
>>
>>The problem in my case is that I had to make several changes and only
>>than make commit.
>>That's why I did the following:
>>
>>1. Turn off Lucene lock (setting the corresponding system variable =
>>false)
>>
>>2. Start the loop (from the first document to the last one to change
>>in the index)
>>
>>2.1. Open IndexReader 
>>2.2. Get a document by its id
>>2.3. Store it as local variable
>>2.4. IndexReader.delete(document id)
>>2.5. IndexReader.close()
>>2.6. Merge new Terms (changes) and old ones in the document I
>>retrieved
>>2.6. Open IndexWriter
>>2.7. Add a new made document
>>
>>3. end of loop
>>
>>4. Waiting for other action ends in my programm I close IndexWriter.
>>
>>The Result:
>>Everything works fine but I had "No tvx file"
>>I really worried about it cause I read what for tvx file...
>>
>>Might anybody explain me what I did wrong?
>>
>>In spite of your answer I did like the following: the way of logging
>>messages
>>
>>This message appeared with the help of System.out.println()
>>
>>Investigating the code of Lucene I found a lot of places of using
>>System.out 
>>
>>I guess it is not a very good solution espessially in so beautiful
>>search/indexing API. 
>>I guess Lucene must have a normal log to write its messages.
>>
>>Thanks in advance...
>>
>>
>>
>>---------------------------------------------------------------------
>>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>>
>>
>>    
>>
>
>
>---------------------------------------------------------------------
>To unsubscribe, e-mail: lucene-user-unsubscribe@jakarta.apache.org
>For additional commands, e-mail: lucene-user-help@jakarta.apache.org
>
>  
>


--------------030902080302050800010400--