convertTo<> { parser {} } can now also defines Char parsing unless a char converter is passed explicitly

Author: Jolanrensen

core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/api/convertTo.kt

@@ -52,7 +52,7 @@ public class ConverterScope(public val fromType: KType, public val toSchema: Col
  * df.convertTo<SomeSchema> {
  *     // defines how to convert Int? -> String
  *     convert<Int?>().with { it?.toString() ?: "No input given" }
- *     // defines how to convert String -> SomeType
+ *     // defines how to convert String/Char -> SomeType
  *     parser { SomeType(it) }
  *     // fill missing column `sum` with expression `a+b`
  *     fill { sum }.with { a + b }
@@ -102,6 +102,10 @@ public fun <T, C> ConvertToFill<T, C>.with(expr: RowExpression<T, C>) {
 
 /**
  * Defines how to convert `String` values into given type [C].
+ *
+ * This method is a shortcut for `convert<String>().with { }`.
+ *
+ * If no converter is defined for `Char` values, this converter will be used for them as well.
  */
 public inline fun <reified C> ConvertSchemaDsl<*>.parser(noinline parser: (String) -> C): Unit =
     convert<String>().with(parser)

core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/impl/api/convertTo.kt

@@ -45,8 +45,10 @@ import org.jetbrains.kotlinx.dataframe.schema.ColumnSchema
 import org.jetbrains.kotlinx.dataframe.schema.DataFrameSchema
 import org.jetbrains.kotlinx.dataframe.size
 import kotlin.reflect.KType
+import kotlin.reflect.full.isSubtypeOf
 import kotlin.reflect.full.withNullability
 import kotlin.reflect.jvm.jvmErasure
+import kotlin.reflect.typeOf
 
 private val logger = KotlinLogging.logger {}
 
@@ -144,6 +146,25 @@ internal fun AnyFrame.convertToImpl(
                         val from = originalColumn.type()
                         val to = targetSchema.type
                         val converter = dsl.getConverter(from, targetSchema)
+                            ?: run {
+                                // Special case for Char columns:
+                                // If there is no explicit Char converter,
+                                // check if we have any converters for String -> target
+                                // if so, we can convert Char -> String -> target
+                                // this allows `parser {}` to work both for Strings and Chars :)
+
+                                if (!from.isSubtypeOf(typeOf<Char?>())) return@run null
+
+                                val stringConverter = dsl.getConverter(
+                                    fromType = typeOf<String>().withNullability(from.isMarkedNullable),
+                                    toSchema = targetSchema,
+                                ) ?: return@run null
+
+                                Converter(
+                                    transform = { stringConverter.transform(this, (it as Char?)?.toString()) },
+                                    skipNulls = stringConverter.skipNulls,
+                                )
+                            }
 
                         val convertedColumn = if (converter != null) {
                             val nullsAllowed = to.isMarkedNullable

core/src/test/kotlin/org/jetbrains/kotlinx/dataframe/api/convertTo.kt

@@ -51,6 +51,38 @@ class ConvertToTests {
         df.convertTo<Schema> { parser { A(it.toInt()) } }
             .single()
             .a.value shouldBe 1
+
+        // shortcut for:
+        df.convertTo<Schema> { convert<String>().with { A(it.toInt()) } }
+            .single()
+            .a.value shouldBe 1
+    }
+
+    @Test
+    fun `convert from char with parser`() {
+        val df = dataFrameOf("a")('1')
+
+        shouldThrow<TypeConverterNotFoundException> {
+            df.convertTo<Schema>()
+        }
+
+        // Char -> String -> Target
+        df.convertTo<Schema> { parser { A(it.toInt()) } }
+            .single()
+            .a.value shouldBe 1
+
+        // shortcut for:
+        df.convertTo<Schema> { convert<String>().with { A(it.toInt()) } }
+            .single()
+            .a.value shouldBe 1
+
+        // Char -> Target
+        df.convertTo<Schema> {
+            parser<A> { error("should not be triggered if convert<Char>() is present") }
+            convert<String>().with<_, A> { error("should not be triggered if convert<Char>() is present") }
+
+            convert<Char>().with { A(it.digitToInt()) }
+        }.single().a.value shouldBe 1
     }
 
     @Test
@@ -335,4 +367,31 @@ class ConvertToTests {
             DataFrame.emptyOf<Entry>(),
         )
     }
+
+    enum class SimpleEnum { A, B }
+
+    @DataSchema
+    interface SchemaWithNullableEnum {
+        val a: SimpleEnum?
+    }
+
+    @Test
+    fun `convert Char to Enum`() {
+        val df = dataFrameOf("a")('A', 'B', null)
+
+        val converted = df.convertTo<SchemaWithNullableEnum>()
+        converted["a"].type() shouldBe typeOf<SimpleEnum?>()
+        converted shouldBe dataFrameOf("a")(SimpleEnum.A, SimpleEnum.B, null)
+    }
+
+    @Test
+    fun `convert Char to Enum custom charParser`() {
+        val df = dataFrameOf("a")('a', 'b', null)
+
+        val converted = df.convertTo<SchemaWithNullableEnum> {
+            parser { SimpleEnum.valueOf(it.uppercase()) }
+        }
+        converted["a"].type() shouldBe typeOf<SimpleEnum?>()
+        converted shouldBe dataFrameOf("a")(SimpleEnum.A, SimpleEnum.B, null)
+    }
 }