core/org.jetbrains.kotlinx.dataframe.api/requireColumn

requireColumn

inline fun <T, reified C> DataFrame<T>.requireColumn(noinline column: ColumnSelector<T, C>): DataFrame<T>

Resolves column in this DataFrame and checks that its runtime type is a subtype of C. Throws if the column can't be resolved or if its type doesn't match.

From the compiler plugin perspective, a new column will appear in the compile-time schema as a result of this operation.

The aim here is to help incrementally migrate workflows to extension properties API.

We recommend considering declaring a DataSchema and use cast or convertTo if you end up with more than a few requireColumn calls.

String API Example:

val repos = DataFrame
    .readCsv("https://raw.githubusercontent.com/Kotlin/dataframe/master/data/jetbrains_repositories.csv")

repos
    .filter { "stargazers_count"<Int>() > 100 }
    .sortByDesc("stargazers_count")
    .select("full_name", "stargazers_count")

Notice how stargazers_count String is repeated three times. We can use requireColumn to add information about this column to the compile-time schema.

Refactored example:

val repos = DataFrame
    .readCsv("https://raw.githubusercontent.com/Kotlin/dataframe/master/data/jetbrains_repositories.csv")
    .requireColumn { "stargazers_count"<Int>() }

repos
    .filter { stargazers_count > 100 }
    .sortByDesc { stargazers_count }
    .select { "full_name" and stargazers_count }

This way code becomes a bit more robust. For example, after .rename { stargazers_count }.into("stars") usages of the stargazers_count property become compile time errors that are easy to spot and update to stars.

val repos = DataFrame
    .readCsv("https://raw.githubusercontent.com/Kotlin/dataframe/master/data/jetbrains_repositories.csv")
    .requireColumn { "stargazers_count"<Int>() }
    .rename { stargazers_count }.into("stars")

repos
    .filter { stars > 100 }
    .sortByDesc { stars }
    .select { "full_name" and stars }

requireColumn can be used to change the type of a known column:

val df = dataFrameOf("name" to columnOf<Any>("Joe"))
val typed = df.requireColumn { name.cast<String>() }
typed.filter { name.length > 2 }

For more information: See requireColumn on the documentation website.