I have downloaded the source code and use File renameTo to do the rename
instead of input/output stream copying. Things are working under Windows. Is it
possible to check OS type and call the correct methods? Anyway, I would put up
a blog when having time so that people can search for the solu
This is actually a known problem, particular to windows OS.
I remember Doug Cuttings has talked about it before.
Inside Lucene, the get-around is to try to rename the file until successful.
The IOException you saw may not affect your indexes' integrity. But I
said so just based on experience, not
Run a small indexer within Eclipse 3.1 under Windows and encounter
rename bug all the time:
java.io.IOException: Cannot rename deleteable.new to deletable
at org.apache.lucene.store.FSDirectory.renameFile(FSDirectory.java:239)
at
org.apache.lucene.index.IndexWriter.writeDeleteableFiles(I
--- Begin Message ---
Run a small indexer within Eclipse 3.1 under Windows and encounter
rename bug all the time:
java.io.IOException: Cannot rename deleteable.new to deletable
at org.apache.lucene.store.FSDirectory.renameFile(FSDirectory.java:239)
at
org.apache.lucene.index.IndexWrite
If you go with one index and distinct fields for the 3 variations you
mentioned, then you can perhaps make use of PerFieldAnalyzerWrapper.
Otis
--- Andrew Boyd <[EMAIL PROTECTED]> wrote:
> Hi All,
> When I first started my project I was creating 3 indexes.
> Standard, Synonym and SoundsLike.
Hi All,
When I first started my project I was creating 3 indexes. Standard, Synonym
and SoundsLike.
Now that the QueryParser has the ability to put multiple tokens in one position
I no longer have to inject the synonyms at index creation time. So I really
don't have to have a seperate index
ok:)
so I do this "filtering" on the hits.
thx
On 7/25/05, Erik Hatcher <[EMAIL PROTECTED]> wrote:
>
> On Jul 25, 2005, at 7:05 AM, Riccardo Daviddi wrote:
> > Hi all,
> > I have to filter my searching in two ways:
> > 1) by score, I want that only docs with score higher than a threshold
> > are
Then your lucene Document can be:
Survey + n Variables,
The idea is just to put everything related to a combined Document, and
search on it.
--
Chris Lu
-
Full-Text Search on Any Database
http://www.dbsight.net
On 7/25/05, Magne Skjeret <[EMAIL PROTECTED]> wrote:
> Chris L
Hi Magne,
Assuming number of variables will not be large, build 2 indexes
Index 1: contents and documents(variable1 + variable2 + variable3...)
Index 2: documents and variables
Example Query: Get "Survey" where "Variable.label" contains "gender"
Step 1) First query would be to "Index 1
Chris Lu wrote:
>Maybe you can have one Document = Survey+Variable,
>and have fields like "abstract", "label".
>
>Just search on the combined documents.
>
>
Hi
A Survey can have 100's of variables. That means that the Survey part is
duplicated N variable times.
Or did I misunderstand you?
Magn
Maybe you can have one Document = Survey+Variable,
and have fields like "abstract", "label".
Just search on the combined documents.
--
Chris Lu
-
Full-Text Search on Any Database
http://www.dbsight.net
On 7/25/05, Magne Skjeret <[EMAIL PROTECTED]> wrote:
> Hi
>
> I have be
Also, be careful with MySql JDBC driver.
Depends on how you use MySql, you could have OutOfMemory errors, which
may not be Lucene or parsers' problem.
--
Chris Lu
-
Full-Text Search on Any Database
http://www.dbsight.net
On 7/25/05, Harini Raghavan <[EMAIL PROTECTED]> wrote:
Hi
I have been reading around a bit on the net to find a solution for my
problem, but no luck yet. I am pretty sure it can be done, but I haven't
found out how yet.
My problem is that I have relations between documents.
E.g
Survey --> Variables
Each survey can have loads of variables, and
I am using org.htmlparser.parserapplications.StringExtractor to parse the
html pages, I guess the OutOfMemory occurs while parsing the large HTML
pages and not while indexing. Sorry about the confusion.
- Original Message -
From: "Erik Hatcher" <[EMAIL PROTECTED]>
To:
Sent: Monday,
What is the point of having empty indexes? I don't know for sure
without trying it myself, but perhaps you need to write at least one
document to the index for it to be a valid index? Don't continue
until Luke is happy with your index.
Erik
On Jul 25, 2005, at 7:39 AM, Daniel Cor
Could you be more specific about where the OutOfMemory error is
happening? Do you have a complete stack trace?
As for maxFieldLength - in my use of Lucene, it is necessary to index
the entire document and not just the first 10,000 or so terms - I set
maxFieldLength to Integer.MAX_VALUE.
On Jul 25, 2005, at 7:05 AM, Riccardo Daviddi wrote:
Hi all,
I have to filter my searching in two ways:
1) by score, I want that only docs with score higher than a threshold
are returned.
2) by num of docs, I want that only, for example, the first 20 docs
are returned.
It's better do it with a
I have a problem searching in 6 index lucene at same time, i founded the
class MultiSearcher but I think I don't use corretly because it show me
this error:
java.io.IOException: Bad file descriptor
at java.io.RandomAccessFile.seek(Native Method)
at
org.apache.lucene.store.FSInputSt
Hi All,
I am using lucene to index large documents(HTML pages). The application is
running on JBoss and MySQL on UNIX. The indexing is throwing OutOfMemory
errors beyond a certain point. I am not sure why this is happening. I am
using the default IndexWriter properties, but the lucene documenta
Hi all,
I have to filter my searching in two ways:
1) by score, I want that only docs with score higher than a threshold
are returned.
2) by num of docs, I want that only, for example, the first 20 docs
are returned.
It's better do it with a queryfilter or search for all and after look
through the
On Jul 24, 2005, at 12:17 PM, Harini Raghavan wrote:
Hi All,
I am trying to add paging functionality while using lucene search.
I have created a PageFilter what takes in the current page num and
the number of records as input and invoking the IndexSearcher
passing the a Boolean Query obj
21 matches
Mail list logo