有人用过Ndatabase对象数据库么，有点问题想咨询下

abutwang 2013-11-21 09:50:37

/// <summary>
/// 添加或更新
/// </summary>
/// <param name="name"></param>
public void StoreOrUpdate(string name)
{
var sport = new Sport(name,new DateTime(2001,1,12));
using (var odb = OdbFactory.Open("test.db"))
{
var query = odb.Query<Sport>();
query.Descend("Name").Constrain(name).Equal();
List<Sport> spotlst = query.Execute<Sport>().ToList<Sport>();
// Store the object
if (spotlst.Count == 0)// 执行查询
{
odb.Store<Sport>(sport);// 保存usermodel对象。
}
else
{
for (int i = 0; i < spotlst.Count; i++)
{
spotlst[i].Dt = DateTime.Now;
odb.Store<Sport>(spotlst[i]);
}
}
}
}

当反复执行此方法进行更新数据时我发现test.db数据文件不断增大

但是问题是我仅仅是update啊
http://download.csdn.net/detail/ruanjianderen/5311887
程序是在csdn上下的很奇怪啊

...全文

195 13 打赏收藏转发到动态举报

写回复

用AI写文章

13 条回复

切换为时间正序

请发表友善的回复…

发表回复

abutwang 2013-11-21

打赏
举报

自己再顶下吧

abutwang 2013-11-21

打赏
举报

引用 1 楼 happy09li 的回复:

http://developer.51cto.com/art/201302/380454.htm 看看相关资料

看过了，你可以试试看他那些资料，如果你每秒钟修改，你会发现数据文件很大很大

熙风 2013-11-21

打赏
举报

http://developer.51cto.com/art/201302/380454.htm 看看相关资料

abutwang 2013-11-21

打赏
举报

/// <summary> /// Store an object with the specific id /// </summary> /// <param name="oid"> </param> /// <param name="plainObject"> </param> private OID InternalStore<T>(OID oid, T plainObject) where T : class { if (GetSession().IsRollbacked()) { throw new OdbRuntimeException( NDatabaseError.OdbHasBeenRollbacked.AddParameter(GetBaseIdentification().ToString())); } if (plainObject == null) throw new OdbRuntimeException(NDatabaseError.OdbCanNotStoreNullObject); var type = typeof (T); if (OdbType.IsNative(type)) { throw new OdbRuntimeException( NDatabaseError.OdbCanNotStoreNativeObjectDirectly.AddParameter(type.FullName).AddParameter( OdbType.GetFromClass(type).Name).AddParameter(type.FullName)); } // first detects if we must perform an insert or an update // If object is in the cache, we must perform an update, else an insert var cache = GetSession().GetCache(); var cacheOid = cache.IdOfInsertingObject(plainObject); if (cacheOid != null) return cacheOid; // throw new ODBRuntimeException("Inserting meta representation of // an object without the object itself is not yet supported"); var mustUpdate = cache.Contains(plainObject); // The introspection callback is used to execute some specific task (like calling trigger, for example) while introspecting the object var callback = _introspectionCallbackForInsert; if (mustUpdate) callback = _introspectionCallbackForUpdate; // Transform the object into an ObjectInfo var nnoi = (NonNativeObjectInfo) _objectIntrospector.GetMetaRepresentation(plainObject, true, null, callback); // During the introspection process, if object is to be updated, then the oid has been set mustUpdate = nnoi.GetOid() != null; return mustUpdate ? _objectWriter.UpdateNonNativeObjectInfo(nnoi, false) : _objectWriter.InsertNonNativeObject(oid, nnoi, true); }

abutwang 2013-11-21

打赏
举报

public override OID Store<T>(OID oid, T plainObject) { if (IsDbClosed) { throw new OdbRuntimeException( NDatabaseError.OdbIsClosed.AddParameter(DbIdentification.Id)); } var newOid = InternalStore(oid, plainObject); GetSession().GetCache().ClearInsertingObjects(); return newOid; }

abutwang 2013-11-21

打赏
举报

引用 9 楼 caozhy 的回复:

[quote=引用 6 楼 abutwang 的回复:] [quote=引用 5 楼 caozhy 的回复:] 看它的源代码就知道。更新数据其实是删除再插入实现的。

那为什么test.db数据文件只大不小呢[/quote] 我没有说清楚。删除其实就是将那块存储标记为无数据。不会将后面的数据移动，填充这个位置。你看下源代码不就知道了。[/quote] 刚在下这个代码呵呵，第一次用这种东西说真的如果这种量级的对象数据库真可以用于实际的话对于小型桌面的应用方便很多啊呵呵