bltoolkit
bltoolkit copied to clipboard
igor-tkachev
Pull Request
I've included all your different advancements and added new functionality since the last time :
1° Added the InsertBatchWithIdentity ( which uses oracle bulk options with OdpDataProvider and Batch insertion sql when using GenericDataProvider with UserQueryText = true). In oracle we reach 1500 insertions (with sequences) per sec. 2° Added the "ContainsExactly" string extension and its linq implementation in order to use the sql fonction "Contains" 3° FullMapping optimisation
"Hello It has been quite some time since we last exchanged some words. I have continued to work on bltoolkit driven by my collegues' and my projects needs and I've tried to get sync with your advancement.
I've added a GenericDataProvider which permits to use any data provider installed on a machine without having a strong link to a specific dataprovider dll (it uses DbProviderFactories concept). I've added a UseQueryText property on DataProvider Base which permits to generate plain text query instead of using command parameters (It's useful for Oracle when some "select queries" are way to fast in plain text). I've added an InsertWithIdentity method on SqlQuery object in order to mentain a similitude with the DbManager class. Some bugs were fixed on JointureAddon which permit now more complex associations.
Would you take a look on these and tell me if it's possible to pull this code into your branch? Thanks! Valeriu"
P.S : I dont have VisualBasic installed so I hope there wont be any compilation errors. But dont hesitate to tell me about.
i'm wondering too ... we are using my fork implementation in a company in about 8 projects in an extensive tested way and he had no problem with it. i should precise that we have mainly oracle and sqlite databases, but we seek for speed optimization in every mean and for practical usages in everyday project, so i think i have some good feedback when developing this fork. but if there's a particular problem with what i've done, i feel eager to repair this. any of you feel free to test this version, to tell me about the new included functionalites.
Is there any way i could help you to integrate this fork? For ex, every time when you do a push on your fork i'm pulling the changes in my fork to facilitate my pull request integration. I'd like to use the nuget packgages for bltoolkit, but i really need the features from my fork.
Such huge pull requests are not reasonable. If you split it into one request per feature without non-bltookit related commits (like SignalR one) I belive your changes it will be integrated much faster.
I agree with you that this pull request is huge. Now. But it wasnt like this all the time. It grew and it grew (pull requests are growing each time you make a new commit) and now it seems relatively hard to integrate. Altough I've done the effort to always integrate the advancement of the main fork, so that should render things easier. What could be done now? Its a bit annoying to have such a split when we can merge things to profit to the entire community...
Maybe this Branch could be merged and the next Version is a unofficial "Beta". And if you don't here any complains you intetgarte it to the main Branch? I'd also would be happy if some of the changes in the Fork got into the main Branch!
Hello jogibear9988.
I've added the FactoryType enumeration which permits to benefit of both the DatabindingMappingSchema and the FullMappingSchema functions.
This option is set into the fullMappingSchema constructor in the factoryType parameter. By default, the factory type value is equal to FactoryType.LazyLoading and so the FullMappingSchema offers only the base features. If you set the parameter value to FactoryType.LazyLoadingWithDataBinding you can then benefit of the options from the 2 schemas.
I'll try to do a better refactoring later in order to have some type inheritage.
FullDataBindingMapping schema class was added and offers both lazy loading and databinding implementation features (among other stuff).
I mention also that the previous commit fixed a bug in the lazy loading of a collection, due to a pull request i've done some time ago.
Hello!
I've removed all the "local" unit tests which couldnt work for you. Could you check please if the pull request is working right now?
Thanks! Valeriu
First of all nice job. It is a very useful framework.
In my poduction environment we are using Oracle databases. After several discussions with Oracle team, there is a case where we have to set the value in the query and not use the the DataParameter : this case is when the column is the partition key. If you use the DataParameter for this field, the explain plan is not optimized.
Thank you in advance to include these optimization.
