What are segments in Lucene?

问题:

What are segments in Lucene?

What are the benefits of segments?

回答1:

The Lucene index is split into smaller chunks called segments. Each segment is its own index. Lucene searches all of them in sequence.

A new segment is created when a new writer is opened and when a writer commits or is closed.

The advantages of using this system are that you never have to modify the files of a segment once it is created. When you are adding new documents in your index, they are added to the next segment. Previous segments are never modified.

Deleting a document is done by simply indicating in a file which document of a segment is deleted, but physically, the document always stays in the segment. Documents in Lucene aren't really updated. What happens is that the previous version of the document is marked as deleted in its original segment and the new version of the document is added to the current segment. This minimizes the chances of corrupting an index by constantly having to modify its content when there are changes. It also allows for easy backup and synchronization of the index across different machines.

However, at some point, Lucene may decide to merge some segments. This operation can also be triggered with an optimize.

回答2:

A segment is very simply a section of the index. The idea is that you can add documents to the index that's currently being served by creating a new segment with only new documents in it. This way, you don't have to go to the expensive trouble of rebuilding your entire index frequently in order to add new documents to the index.

回答3:

The segment benefits have been answered already by others. I will include an ascii diagram of a Lucene Index.

Lucene Segment

A Lucene segment is part of an Index. Each segment is composed of several index files. If you look inside any of these files, you will see that it holds 1 or more Lucene documents.

+- Index 5 ------------------------------------------+
|                                                    |
|  +- Segment _0 ---------------------------------+  |
|  |                                              |  |
|  |  +- file 1 -------------------------------+  |  |
|  |  |                                        |  |  |
|  |  | +- L.Doc1-+  +- L.Doc2-+  +- L.Doc3-+  |  |  |
|  |  | |         |  |         |  |         |  |  |  |
|  |  | | field 1 |  | field 1 |  | field 1 |  |  |  |
|  |  | | field 2 |  | field 2 |  | field 2 |  |  |  |
|  |  | | field 3 |  | field 3 |  | field 3 |  |  |  |
|  |  | |         |  |         |  |         |  |  |  |
|  |  | +---------+  +---------+  +---------+  |  |  |
|  |  |                                        |  |  |
|  |  +----------------------------------------+  |  |
|  |                                              |  |
|  |                                              |  |
|  |  +- file 2 -------------------------------+  |  |
|  |  |                                        |  |  |
|  |  | +- L.Doc4-+  +- L.Doc5-+  +- L.Doc6-+  |  |  |
|  |  | |         |  |         |  |         |  |  |  |
|  |  | | field 1 |  | field 1 |  | field 1 |  |  |  |
|  |  | | field 2 |  | field 2 |  | field 2 |  |  |  |
|  |  | | field 3 |  | field 3 |  | field 3 |  |  |  |
|  |  | |         |  |         |  |         |  |  |  |
|  |  | +---------+  +---------+  +---------+  |  |  |
|  |  |                                        |  |  |
|  |  +----------------------------------------+  |  |
|  |                                              |  |
|  +----------------------------------------------+  |
|                                                    |
|  +- Segment _1 (optional) ----------------------+  |
|  |                                              |  |
|  +----------------------------------------------+  |
+----------------------------------------------------+

Reference

Lucene in Action Second Edition - July 2010 - Manning Publication