modeldb | Open Source ML Model Versioning , Metadata , and Experiment | Machine Learning library

It looks like in the following line:

max_date = select(max(cls.prediction_model.c.created_datetime)).\

you're using Python's max function, rather than passing the sql max function. Can you try using max from sqlalchemy instead? It sits in the func module. So write:

max_date = select([func.max(cls.prediction_model.c.created_datetime)]).\

Note that you also need to use a list in your select query, hence the added [ ].

On another note, if the ForeignKey relationships have been defined correctly in your table definitions, you don't need to tell sqlalchemy explicitely how to do the join. So in that case you could get away with something like:

Using the great suggestion from @teunbrand, this could help:

As stays in the message split will be not supported in the linq to sql and if you want to get result without additional methods, using Contains() should give you required result

Welcome to programming,

you should echo your SQL query like below to evaluate what kind of query your input generates.

echo 'SELECT ID, modelName, datecreated FROM modeldb ORDER BY "' . $name . '"'

you will find that the " are not required at all.

one rule of programming is that never trust the user input. here you are injecting the data which you received from $_POST so you should validate your input like

My first advice is to do bulk inserts in the DBMS, not in code. Doing them via code is only prone to add addtional issues.

As far as parsing the .xlsx files go, the OleDB driver is propably unessesary. There are a few basic rules for working with office formats:

• if you can limit it to the new ones (.xlsx), you can use the OpenXML SDK. Or any of the Wrappers people made around it. Or even just the .ZipArchive and XMLReader classes.
• if you need to support the old formats (.xls) too, you got to use the (t)rusty Office COM Interop. This has all the usual issue of COM Interop, and aditionally needs office installed and a Interactive session
• for any given Display Technology and Problem, there might be a 3rd option. But those are few and far in between. As we always got the Interop to fall back on, we never developed a complete Office processing class like so many other languages have.

I would put OlebDB in that last category - a rare and very specific solution.

It always advise for using the first option.

And Officer COM Interop should be silently burried, with the old formats being removed from the file format options. Considering this is a WebApplication that will likely run as services, you will not get the nesseary interactive session anyway.

Of course accepting .csv is also an option. And indeed Excel has full .CSV support.

It looks like you didn't update list of modelDB's inside adapter after the parsing of a network call result.

Are you using Code First? If so I think you need to generate migrations. In visual studio go to Package Manager Console and run this commands:

• Add-Migration "modelClassName"
• Update-Database –Verbose

For more information refer to this link: https://msdn.microsoft.com/en-us/library/jj591621(v=vs.113).aspx

You are missing the set; in the field ID.

Unfortunately there is no out of the box way to do that. But you can create a Custom Naming Convention like this:

The main issue is that you are initializing the reference navigation property (aDepartment = new Department();) which is confusing the EF infrastructure and prevents it working properly. Initializing collection navigation properties is ok (although not mandatory), but never do that for reference navigation properties.

Removing the above line from the Subject constructor will fix the issue. Additionally, although it somehow works in your case (surprisingly to me), better follow the naming conventions and don't call navigation property aDepartment, but simply Department to line up with the DepartmentID FK property.

Here is the corrected working version of the Subject entity: