# ColumnSelector An operation that applies an estimator to a selection of columns. ``` struct ColumnSelector where Estimator : Estimator, Estimator.Transformer.Input == UnwrappedInput? ``` ## Overview This estimator applies a non-tabular estimator to a selection of columns. Here’s an example of normalizing numeric values within each column using a [`StandardScaler`](/documentation/CreateMLComponents/StandardScaler): ``` let numericalScaling = ColumnSelector( columns: ["volume", "price"], estimator: NumericImputer(.mean) .appending(StandardScaler()) ) ``` In most cases, an inputer must handle missing values. ## Topics ### Creating the selection [`init(columns: [String], estimator: Estimator)`](/documentation/CreateMLComponents/ColumnSelector/init(columns:estimator:)) Creates a select operation with an estimator. [`init(ColumnSelection, estimator: Estimator)`](/documentation/CreateMLComponents/ColumnSelector/init(_:estimator:)) Creates a select operation with an estimator. [`init(ColumnSelection, transformer: T)`](/documentation/CreateMLComponents/ColumnSelector/init(_:transformer:)) Creates a select operation with a transformer. ### Getting the properties [`var columnSelection: ColumnSelection`](/documentation/CreateMLComponents/ColumnSelector/columnSelection) The column selection strategy. [`var estimator: Estimator`](/documentation/CreateMLComponents/ColumnSelector/estimator) The estimator to use on each column. ### Encoding and decoding [`func encode(ColumnSelector.Transformer, to: inout any EstimatorEncoder) throws`](/documentation/CreateMLComponents/ColumnSelector/encode(_:to:)) Encodes a fitted transformer. [`func decode(from: inout any EstimatorDecoder) throws -> ColumnSelector.Transformer`](/documentation/CreateMLComponents/ColumnSelector/decode(from:)) Decodes a previously fitted transformer. ### Fitting a transformer [`func fitted(to: DataFrame, eventHandler: EventHandler?) async throws -> ColumnSelector.Transformer`](/documentation/CreateMLComponents/ColumnSelector/fitted(to:eventHandler:)) Fits a transformer to a data frame [`typealias Input`](/documentation/CreateMLComponents/ColumnSelector/Input) [`typealias Output`](/documentation/CreateMLComponents/ColumnSelector/Output) [`protocol Transformer`](/documentation/CreateMLComponents/Transformer) A transformer that takes an input and produces an output. ### Default Implementations [UpdatableTabularEstimator Implementations](/documentation/CreateMLComponents/ColumnSelector/UpdatableTabularEstimator-Implementations) ## Relationships ### Conforms To [`SendableMetatype`](/documentation/Swift/SendableMetatype) [`Sendable`](/documentation/Swift/Sendable) [`Copyable`](/documentation/Swift/Copyable) [`UpdatableTabularEstimator`](/documentation/CreateMLComponents/UpdatableTabularEstimator) [`Escapable`](/documentation/Swift/Escapable) [`TabularEstimator`](/documentation/CreateMLComponents/TabularEstimator) --- Copyright © 2026 Apple Inc. All rights reserved. | [Terms of Use](https://www.apple.com/legal/internet-services/terms/site.html) | [Privacy Policy](https://www.apple.com/privacy/privacy-policy)