在后台线程中使用 MagicalRecord 和 NSXMLParser [英] Using MagicalRecord and NSXMLParser in a background thread

问题描述

我正在使用 NSXMLParser 来处理需要在我的核心数据模型中结束的大型复杂 XML 文件.为此，我遵循 Conway 和 Hillegass 在 iOS 书中讨论的设计模式，每次到达新节点并动态创建新节点时，NSXMLParser 的委托都会更改.为了创建实体，我在解析过程中为每个新节点使用 MagicalRecord 的 MR_createEntity 以设置我的数据模型.解析完成后我调用

I'm using an NSXMLParser to process a large and complex XML file that needs to end up in my Core Data model. To do this I am following the design pattern as discussed in the iOS book by Conway and Hillegass, where the delegate of the NSXMLParser is changed every time a new node is reached and new nodes are created on the fly. To create the entities, I am using MagicalRecord's MR_createEntity for each new node during the parsing to set up my data model. After the parsing is finished I call

[[NSManagedObjectContext MR_defaultContext] MR_saveToPersistentStoreAndWait]; 

将新实体保存到我的核心数据存储中.

to save the new entities to my Core Data store.

这一切都很好，直到我最近决定将解析移到后台线程中.我注意到 XML 数据并不总是被导入.因此，经过一番谷歌搜索后，我发现 MR_createEntity 不应该用于后台线程(链接:https://github.com/magicalpanda/MagicalRecord/issues/298).

This all works fine, until I recently decided to move the parsing into a background thread. And I noticed the XML data was not always imported. So after some Googling I found that MR_createEntity should not be used on a background thread (link: https://github.com/magicalpanda/MagicalRecord/issues/298).

那怎么办?我看到了两种可能的解决方案:

So what to do? I see two possible solutions:

1. 在解析过程中，只需为所有节点创建一个基于 Foundation 的结构，即 NSDictionaries 和 NSArrays 的混合体.这可以在后台线程上完成.完成后，我返回主线程并使用 MR_importValuesForKeysWithObject 或类似的东西将我的数据保存到 Core Data 中(如下所述:http://www.cimgf.com/2012/05/29/importing-data-made-easy/).但这是否适用于深度嵌套的字典和数组的组合?

1. During the parsing, just create a Foundation based structure of all the nodes, a mixture of NSDictionaries and NSArrays. This can be done on a background thread. Once done, I go back to the main thread and save my data into Core Data using MR_importValuesForKeysWithObject or something similar (as described here: http://www.cimgf.com/2012/05/29/importing-data-made-easy/). But will that work with combinations of deeply nested dictionaries and arrays?

在解析过程中，我没有调用MR_CreateEntity，而是使用

During the parsing, instead of calling MR_CreateEntity, I use

[MagicalRecord saveInBackgroundWithBlock:^(NSManagedObjectContext *localContext)
{
    MyEntity *entity = [MyEntity MR_createInContext:localContext];
}];

每当我遇到上面 github 链接中建议的新节点时.

every time I come upon a new node as suggested in the github link above.

那么在我开始削减和修改我的代码之前，应该怎么做?也许还有另一种方法?

So before I start cutting up and modifying my code, what would be the way to go? Maybe there is another approach?

推荐答案

如果我正确理解 MagicalRecord 源代码，你只需要一个单saveInBackgroundWithBlock:completion: 调用:

If I understand the MagicalRecord source code correctly, you need only a single saveInBackgroundWithBlock:completion: call:

[MagicalRecord saveInBackgroundWithBlock:^(NSManagedObjectContext *localContext) {
      // ... parse XML ...
      // ... create entities with:
      MyEntity *entity = [MyEntity MR_createInContext:localContext];
      // ...
} completion:^{
    NSLog("Import finished");
}];

这个

• 创建一个临时的背景上下文，
• 执行与该上下文关联的后台队列中的第一个块，
• 调用 MR_saveToPersistentStoreAndWait 将临时上下文保存到主上下文和将主上下文保存到持久化核心，
• 最后执行完成块.

• creates a temporary background context,
• executes the first block on the background queue associated with that context,
• calls MR_saveToPersistentStoreAndWait to save the temporary context to the main context and save the main context to the persistent core,
• and finally executes the completion block.

这篇关于在后台线程中使用 MagicalRecord 和 NSXMLParser的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！