The source data for my index is already in standard UTF-8 and available as a
simple byte array. I need to do some simple tokenization of the data (check
for whitespace and special characters that control position increment). What
is the most efficient way to index this data and avoid unnecessary
conversions to/from java Strings or char arrays? Looking at DocumentsWriter,
I see that all terms are eventually converted to char arrays and written in
modified-UTF-8, so there doesn't seem to be much advantage to having the
source data in standard UTF-8.
Peter
On 2/14/07, Chris Hostetter <[EMAIL PROTECTED]> wrote:
Internally Lucene deals with pure Java Strings; when writing those strings
to and reading those strings back from disk, Lucene allways uses the stock
Java "modified UTF-8" format, regardless of what your file.encoding
system property may be.
typcially when people have encoding problems in their lucene applications,
the origin of hte problem is in the way they fetch the data before
indexing it ... if you can make a String object, and System.out.println
that string and see what you expect, then handing that string to Lucene as
a field value should work fine.
what exactly is the "value" object you are calling getBytes on? ... if
it's another String, then you've already got serious problems -- i can't
imagine any situation where fetching the bytes from a String in one
charset and using those bytes to construct another string (either in a
different charset, or in the system default charset) would make any sense
at all.
wherever your original binary data is coming from (files on disk, network
socket, etcc...) that's when you should be converting those bytes into
chars using whatever charset you know those bytes represent.
: Date: Wed, 14 Feb 2007 09:16:58 +0330
: From: Mohammad Norouzi <[EMAIL PROTECTED]>
: Reply-To: java-user@lucene.apache.org
: To: java-user@lucene.apache.org
: Subject: encoding question.
:
: Hi
: I want to index data with utf-8 encoding, so when adding field to a
document
: I am using the code new String(value.getBytes("utf-8"))
: in the other hand, when I am going to search I was using the same
snippet
: code to convert to utf-8 but it did not work so finally I found
somewhere
: that had been said to use new String(valueToSearch.getBytes
("cp1252"),"UTF8")
: and it worked fine but I still has some problem.
: first, some characters are weird when I get result from lucene, It seems
it
: is in cp1252 encoding.
: second, if the java environment property "file.encoding" not been cp1252
the
: result is completely in incorrect encoding. so I must change this
property
: using System.setProperty("file.encoding","cp1252")
:
: is lucene neglect my utf-8 encoding and proceed indexing data using
cp1252?
: how can I correct weird characters I received by searching?
:
: Thank you very much in advance.
: --
: Regards,
: Mohammad
:
-Hoss
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
