다음을 통해 공유


TransformExtensionsCatalog.SelectColumns 메서드

정의

오버로드

SelectColumns(TransformsCatalog, String[])

Create a ColumnSelectingEstimator, which keeps a given list of columns in an IDataView and drops the others.

SelectColumns(TransformsCatalog, String[], Boolean)

Create a ColumnSelectingEstimator, which keeps a given list of columns in an IDataView and drops the others.

SelectColumns(TransformsCatalog, String[])

Create a ColumnSelectingEstimator, which keeps a given list of columns in an IDataView and drops the others.

public static Microsoft.ML.Transforms.ColumnSelectingEstimator SelectColumns (this Microsoft.ML.TransformsCatalog catalog, params string[] columnNames);
static member SelectColumns : Microsoft.ML.TransformsCatalog * string[] -> Microsoft.ML.Transforms.ColumnSelectingEstimator
<Extension()>
Public Function SelectColumns (catalog As TransformsCatalog, ParamArray columnNames As String()) As ColumnSelectingEstimator

매개 변수

catalog
TransformsCatalog

변환의 카탈로그입니다.

columnNames
String[]

유지할 열 이름의 배열입니다.

반환

예제

using System;
using System.Collections.Generic;
using Microsoft.ML;

namespace Samples.Dynamic
{
    public static class SelectColumns
    {
        public static void Example()
        {
            // Create a new ML context, for ML.NET operations. It can be used for
            // exception tracking and logging, as well as the source of randomness.
            var mlContext = new MLContext();

            // Create a small dataset as an IEnumerable.
            var samples = new List<InputData>()
            {
                new InputData(){ Age = 21, Gender = "Male", Education = "BS",
                    ExtraColumn = 1 },

                new InputData(){ Age = 23, Gender = "Female", Education = "MBA",
                    ExtraColumn = 2 },

                new InputData(){ Age = 28, Gender = "Male", Education = "PhD",
                    ExtraColumn = 3 },

                new InputData(){ Age = 22, Gender = "Male", Education = "BS",
                    ExtraColumn = 4 },

                new InputData(){ Age = 23, Gender = "Female", Education = "MS",
                    ExtraColumn = 5 },

                new InputData(){ Age = 27, Gender = "Female", Education = "PhD",
                    ExtraColumn = 6 },
            };

            // Convert training data to IDataView.
            var dataview = mlContext.Data.LoadFromEnumerable(samples);

            // Select a subset of columns to keep.
            var pipeline = mlContext.Transforms.SelectColumns("Age", "Education");

            // Now we can transform the data and look at the output to confirm the
            // behavior of SelectColumns. Don't forget that this operation doesn't
            // actually evaluate data until we read the data below, as
            // transformations are lazy in ML.NET.
            var transformedData = pipeline.Fit(dataview).Transform(dataview);

            // Print the number of columns in the schema
            Console.WriteLine($"There are {transformedData.Schema.Count} columns" +
                $" in the dataset.");

            // Expected output:
            //  There are 2 columns in the dataset.

            // We can extract the newly created column as an IEnumerable of
            // TransformedData, the class we define below.
            var rowEnumerable = mlContext.Data.CreateEnumerable<TransformedData>(
                transformedData, reuseRowObject: false);

            // And finally, we can write out the rows of the dataset, looking at the
            // columns of interest.
            Console.WriteLine($"Age and Educations columns obtained " +
                $"post-transformation.");

            foreach (var row in rowEnumerable)
                Console.WriteLine($"Age: {row.Age} Education: {row.Education}");

            // Expected output:
            //  Age and Educations columns obtained post-transformation.
            //  Age: 21 Education: BS
            //  Age: 23 Education: MBA
            //  Age: 28 Education: PhD
            //  Age: 22 Education: BS
            //  Age: 23 Education: MS
            //  Age: 27 Education: PhD
        }

        private class InputData
        {
            public int Age { get; set; }
            public string Gender { get; set; }
            public string Education { get; set; }
            public float ExtraColumn { get; set; }
        }

        private class TransformedData
        {
            public int Age { get; set; }
            public string Education { get; set; }
        }
    }
}

적용 대상

SelectColumns(TransformsCatalog, String[], Boolean)

Create a ColumnSelectingEstimator, which keeps a given list of columns in an IDataView and drops the others.

public static Microsoft.ML.Transforms.ColumnSelectingEstimator SelectColumns (this Microsoft.ML.TransformsCatalog catalog, string[] columnNames, bool keepHidden);
static member SelectColumns : Microsoft.ML.TransformsCatalog * string[] * bool -> Microsoft.ML.Transforms.ColumnSelectingEstimator
<Extension()>
Public Function SelectColumns (catalog As TransformsCatalog, columnNames As String(), keepHidden As Boolean) As ColumnSelectingEstimator

매개 변수

catalog
TransformsCatalog

변환의 카탈로그입니다.

columnNames
String[]

유지할 열 이름의 배열입니다.

keepHidden
Boolean

숨겨진 true 열을 유지하고 숨겨진 열을 false 제거합니다. 디버깅을 위해 파이프라인의 입력이 파이프라인의 출력에 매핑되는 방식을 이해해야 하는 경우 숨겨진 열을 삭제하는 대신 유지하는 것이 좋습니다.

반환

예제

using System;
using System.Collections.Generic;
using Microsoft.ML;

namespace Samples.Dynamic
{
    public static class SelectColumns
    {
        public static void Example()
        {
            // Create a new ML context, for ML.NET operations. It can be used for
            // exception tracking and logging, as well as the source of randomness.
            var mlContext = new MLContext();

            // Create a small dataset as an IEnumerable.
            var samples = new List<InputData>()
            {
                new InputData(){ Age = 21, Gender = "Male", Education = "BS",
                    ExtraColumn = 1 },

                new InputData(){ Age = 23, Gender = "Female", Education = "MBA",
                    ExtraColumn = 2 },

                new InputData(){ Age = 28, Gender = "Male", Education = "PhD",
                    ExtraColumn = 3 },

                new InputData(){ Age = 22, Gender = "Male", Education = "BS",
                    ExtraColumn = 4 },

                new InputData(){ Age = 23, Gender = "Female", Education = "MS",
                    ExtraColumn = 5 },

                new InputData(){ Age = 27, Gender = "Female", Education = "PhD",
                    ExtraColumn = 6 },
            };

            // Convert training data to IDataView.
            var dataview = mlContext.Data.LoadFromEnumerable(samples);

            // Select a subset of columns to keep.
            var pipeline = mlContext.Transforms.SelectColumns("Age", "Education");

            // Now we can transform the data and look at the output to confirm the
            // behavior of SelectColumns. Don't forget that this operation doesn't
            // actually evaluate data until we read the data below, as
            // transformations are lazy in ML.NET.
            var transformedData = pipeline.Fit(dataview).Transform(dataview);

            // Print the number of columns in the schema
            Console.WriteLine($"There are {transformedData.Schema.Count} columns" +
                $" in the dataset.");

            // Expected output:
            //  There are 2 columns in the dataset.

            // We can extract the newly created column as an IEnumerable of
            // TransformedData, the class we define below.
            var rowEnumerable = mlContext.Data.CreateEnumerable<TransformedData>(
                transformedData, reuseRowObject: false);

            // And finally, we can write out the rows of the dataset, looking at the
            // columns of interest.
            Console.WriteLine($"Age and Educations columns obtained " +
                $"post-transformation.");

            foreach (var row in rowEnumerable)
                Console.WriteLine($"Age: {row.Age} Education: {row.Education}");

            // Expected output:
            //  Age and Educations columns obtained post-transformation.
            //  Age: 21 Education: BS
            //  Age: 23 Education: MBA
            //  Age: 28 Education: PhD
            //  Age: 22 Education: BS
            //  Age: 23 Education: MS
            //  Age: 27 Education: PhD
        }

        private class InputData
        {
            public int Age { get; set; }
            public string Gender { get; set; }
            public string Education { get; set; }
            public float ExtraColumn { get; set; }
        }

        private class TransformedData
        {
            public int Age { get; set; }
            public string Education { get; set; }
        }
    }
}

적용 대상