A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.
Python artificial-intelligencedagdata-sciencedata-visualizationdataflowdeveloper-toolsmachine-learningnotebookspipelinepythonreactivesqlweb-app
/* Copyright 2026 Marimo. All rights reserved. */
import type { FieldTypesWithExternalType } from "@/components/data-table/types";
import type { ColumnDataTypes, ColumnId } from "../types";
/**
* Get the effective columns at a given transform step.
*
* @param columns - Original columns (fallback)
* @param columnTypesPerStep - Column types at each step from backend
* - Index 0 = original columns
* - Index N = columns after transform N-1
* @param selectedTransform - Index of the currently selected transform
* @returns The effective columns before the selected transform
*/
export function getEffectiveColumns(
columns: ColumnDataTypes,
columnTypesPerStep: FieldTypesWithExternalType[] | undefined,
stepIndex = 0,
): ColumnDataTypes {
// If no columnTypesPerStep, fall back to original columns
if (!columnTypesPerStep || columnTypesPerStep.length === 0) {
return columns;
}
// columnTypesPerStep[0] = original columns
// columnTypesPerStep[N] = columns after transform N-1
// For the selected transform, we want columns BEFORE it, so index = selectedTransform
const safeIndex = Math.min(stepIndex, columnTypesPerStep.length - 1);
const fieldTypes = columnTypesPerStep[safeIndex];
if (!fieldTypes) {
return columns;
}
const newColumns = new Map<ColumnId, string>();
for (const [name, [dataType]] of fieldTypes) {
newColumns.set(name as ColumnId, dataType);
}
return newColumns;
}