machinelearning icon indicating copy to clipboard operation
machinelearning copied to clipboard

Added ADO.NET importing/exporting functionality to DataFrame

Open andrei-faber opened this issue 4 years ago • 9 comments

See https://github.com/dotnet/machinelearning/issues/5972 This universal integration with ADO.NET allows importing and exporting data to a large number of SQL-compatible databases. In addition to this, this PR adds methods to load data from any IEnumerable<T> collections and to export data to System.Data.DataTable

andrei-faber avatar Oct 14 '21 16:10 andrei-faber

CLA assistant check
All CLA requirements met.

dnfadmin avatar Oct 14 '21 16:10 dnfadmin

It seems that build fails because I added a reference to the System.Data.SQLite NuGet package (it's only used for integration tests), and the build doesn't use public NuGet repos; I wonder if it's possible to add System.Data.SQLite to the private repo?

andrei-faber avatar Oct 14 '21 16:10 andrei-faber

@andrei-faber I believe that should be just fine. Let me talk to some people on my end and I'll get back to you.

michaelgsharp avatar Oct 14 '21 17:10 michaelgsharp

/azp run

michaelgsharp avatar Oct 20 '21 17:10 michaelgsharp

Azure Pipelines successfully started running 2 pipeline(s).

azure-pipelines[bot] avatar Oct 20 '21 17:10 azure-pipelines[bot]

@andrei-faber The NuGet package has been added and I have requeued the pipelines.

michaelgsharp avatar Oct 20 '21 17:10 michaelgsharp

@michaelgsharp some strange version conflict in there. Maybe it's better to use libraries with the same version number for System.Data.SQLite.Core in Microsoft.ML.Tests.csproj and System.Data.SQLite in Microsoft.Data.Analysis.Tests.csproj For instance use System.Data.SQLite 1.0.112.2 (if it's possible to get it)

andrei-faber avatar Oct 21 '21 20:10 andrei-faber

@andrei-faber So part of the issue is that we are still building for .net core 2, which the sql package doesn't support. I am in the process of removing that, should be done by next week (have a couple of things do to first), that should resolve this issue. So we should be able to get this in next week.

I did get the version stuff figured out, I didn't get the package moved correctly.

michaelgsharp avatar Oct 28 '21 19:10 michaelgsharp

@michaelgsharp great, thanks. Is there anything I can do to help?

andrei-faber avatar Oct 28 '21 20:10 andrei-faber

Codecov Report

Merging #5975 (cf05e79) into main (ff3b1b9) will increase coverage by 0.01%. The diff coverage is 86.36%.

Additional details and impacted files
@@            Coverage Diff             @@
##             main    #5975      +/-   ##
==========================================
+ Coverage   68.65%   68.67%   +0.01%     
==========================================
  Files        1202     1203       +1     
  Lines      250769   250974     +205     
  Branches    26190    26209      +19     
==========================================
+ Hits       172166   172344     +178     
- Misses      71785    71805      +20     
- Partials     6818     6825       +7     
Flag Coverage Δ
Debug 68.67% <86.36%> (+0.01%) :arrow_up:
production 63.16% <79.31%> (+0.01%) :arrow_up:
test 88.87% <100.00%> (+0.01%) :arrow_up:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files Coverage Δ
src/Microsoft.Data.Analysis/Extensions.cs 73.33% <73.33%> (ø)
src/Microsoft.Data.Analysis/DataFrame.IO.cs 80.76% <80.00%> (+0.88%) :arrow_up:
...Microsoft.Data.Analysis.Tests/DataFrame.IOTests.cs 98.97% <100.00%> (+0.07%) :arrow_up:

... and 5 files with indirect coverage changes

codecov[bot] avatar May 06 '23 07:05 codecov[bot]