(ZODB データベース)¶
Description
Plone uses ZODB database to store its data. ZODB can act independently in-process, clustered over network or over another database engine, like SQL.
local
はじめに¶
Plone uses ZODB database. ZODB happily eats any Python object with any attributes - there is no need to write “database schemas” as it is with SQL based systems.
Classes inherit for persistent.Persistent will have all their attributes written to database. Lists and dictionaries will be automatically converted to persistent versions.
This chapter is about basics of ZODB, working with ZODB database directly, like tuning database settings.
Object-oriented database constraints¶
ZODB is object-oriented database. It makes very easy to store different kind of contentish data in a nested tree, supporting subclassing (something which SQL often does poorly).
Since the database is object oriented and objects are defined in Python code, you always need accomppaning Python source code to read what’s inside ZODB. This might feel akward at first, but you need running MySQL to read what’s inside MySQL files stored on your disk and so on...
警告
ZODB database is not usable without the Python source code used to create the data. The data is not readable using any of standard tools and there exist little tools to deal with the raw data. The way to access Plone data is running the Plone itself and performing queries through it.
警告
Since correct source code is needed to read ZODB data, this poses a problem for versioning. Even if you use the correct add-on product with proper source code, if the source code version if wrong, it might not work. Data model attributes might be added, modified or deleted between source code revisions, making data operations on the existing database fail by raising Python Exceptions (AttributeError, KeyError).
To workaround the ZODB interoperatibility problems, products like ore.contentmirror exist to duplicate Plone content data to read-only SQL database.
Query and searching¶
ZODB does not provide query services as is i.e. there is no SELECT statement.
Plone provides cataloging service for this purpose.
This gives some benefits
- You define yourself how data is indexed
- The backend to perform queries is flexible - you can plug-in custom indexes
- portal_catalog default catalog is used to all content items to provide basic CMS functionality easily
- You can have optimized catalogs for specialized data (e.g. reference look-ups using reference_catalog)
Data model¶
There is no hardwired way for desceribe data in ZODB database.
There are currently three primary ways to define data models in Plone
- Using Archetypes content type subsystem (all Plone 3 content)
- Using zope.schema package (modern way)
- Not defining the model, but relying
Read about zope.schema how to define a model for the adta to be stored in ZODB database.
ZODB versions¶
Plone uses ZODB database. The default ZODB version with Plone 3.x is ZODB 3.7.x. ZODB 3.8.x is not officially supported, but has been reported to work to work with Plone 3. ZODB 3.9.x is to be used with Plone 4.
ZODB tips and tricks¶
Please see
Packing the database offline¶
Example how to pack a copy of Data.fs in offline using Python snippet:
import time
import ZODB.FileStorage
import ZODB.serialize
storage=ZODB.FileStorage.FileStorage('/tmp/Data.fs.copy')
storage.pack(time.time(),ZODB.serialize.referencesf)
