init research

dataframe/docs/StardustDocs/topics/unfold.md

@@ -0,0 +1,60 @@
+[//]: # (title: unfold)
+<!---IMPORT org.jetbrains.kotlinx.dataframe.samples.api.Modify-->
+
+Returns [`DataFrame`](DataFrame.md) in which selected data columns are converted to [`ColumnGroup`](DataColumn.md#columngroup) / [`FrameColumn`](DataColumn.md#framecolumn) according to
+the values of the properties of the objects.
+
+It's a special case of [convert](convert.md) operation. 
+
+This operation is useful when: 
+1. you use a library API that gives you class instances
+2. you do not want to or cannot annotate classes with [`@DataSchema`](schemas.md)
+
+See [column selectors](ColumnSelectors.md) for how to select the columns for this operation.
+
+### Library API
+
+<!---FUN convertToColumnGroupUseCase-->
+
+```kotlin
+class RepositoryInfo(val data: Any)
+
+fun downloadRepositoryInfo(url: String) = RepositoryInfo("fancy response from the API")
+```
+
+<!---END-->
+
+Consider you have an existing [`DataFrame`](DataFrame.md) with some URLs, arguments for an API call. 
+
+<!---FUN convertToColumnGroupData-->
+
+```kotlin
+val interestingRepos = dataFrameOf("name", "url")(
+    "dataframe", "/dataframe",
+    "kotlin", "/kotlin",
+)
+
+val initialData = interestingRepos
+    .add("response") { downloadRepositoryInfo("url"<String>()) }
+```
+
+<!---END-->
+
+Using unfold, you can convert `response` to a [`ColumnGroup`](DataColumn.md#columngroup) and use rich [modify](modify.md) capabilities.
+
+<!---FUN convertToColumnGroup-->
+
+```kotlin
+val df = initialData.unfold("response")
+```
+
+<!---END-->
+
+<!---FUN convertToColumnGroupBenefits-->
+
+```kotlin
+df.move { response.data }.toTop()
+df.rename { response.data }.into("description")
+```
+
+<!---END-->