Merge pull request #835 from Kotlin/empty-csv-fix

Empty csv fix

Author: Jolanrensen

core/src/main/kotlin/org/jetbrains/kotlinx/dataframe/io/csv.kt

@@ -345,7 +345,7 @@ public fun DataFrame.Companion.readDelim(
     }
 
     val columnNames = csvParser.headerNames.takeIf { it.isNotEmpty() }
-        ?: (1..records[0].count()).map { index -> "X$index" }
+        ?: (1..(records.firstOrNull()?.count() ?: 0)).map { index -> "X$index" }
 
     val generator = ColumnNameGenerator()
     val uniqueNames = columnNames.map { generator.addUnique(it) }

core/src/test/kotlin/org/jetbrains/kotlinx/dataframe/io/CsvTests.kt

@@ -13,6 +13,7 @@ import org.jetbrains.kotlinx.dataframe.api.dataFrameOf
 import org.jetbrains.kotlinx.dataframe.api.group
 import org.jetbrains.kotlinx.dataframe.api.groupBy
 import org.jetbrains.kotlinx.dataframe.api.into
+import org.jetbrains.kotlinx.dataframe.api.isEmpty
 import org.jetbrains.kotlinx.dataframe.api.schema
 import org.jetbrains.kotlinx.dataframe.api.toStr
 import org.jetbrains.kotlinx.dataframe.ncol
@@ -282,6 +283,40 @@ class CsvTests {
         df.columnNames() shouldBe listOf("Column1", "Column2")
     }
 
+    @Test
+    fun `read empty delimStr or CSV`() {
+        val emptyDelimStr = DataFrame.readDelimStr("")
+        emptyDelimStr shouldBe DataFrame.empty()
+
+        val emptyDelimFile = DataFrame.readDelim(File.createTempFile("empty", "csv").reader())
+        emptyDelimFile shouldBe DataFrame.empty()
+
+        val emptyCsvFile = DataFrame.readCSV(File.createTempFile("empty", "csv"))
+        emptyCsvFile shouldBe DataFrame.empty()
+
+        val emptyCsvFileManualHeader = DataFrame.readCSV(
+            file = File.createTempFile("empty", "csv"),
+            header = listOf("a", "b", "c"),
+        )
+        emptyCsvFileManualHeader.apply {
+            isEmpty() shouldBe true
+            columnNames() shouldBe listOf("a", "b", "c")
+            columnTypes() shouldBe listOf(typeOf<String>(), typeOf<String>(), typeOf<String>())
+        }
+
+        val emptyCsvFileWithHeader = DataFrame.readCSV(
+            file = File.createTempFile("empty", "csv").also { it.writeText("a,b,c") },
+        )
+        emptyCsvFileWithHeader.apply {
+            isEmpty() shouldBe true
+            columnNames() shouldBe listOf("a", "b", "c")
+            columnTypes() shouldBe listOf(typeOf<String>(), typeOf<String>(), typeOf<String>())
+        }
+
+        val emptyTsvStr = DataFrame.readTSV(File.createTempFile("empty", "tsv"))
+        emptyTsvStr shouldBe DataFrame.empty()
+    }
+
     companion object {
         private val simpleCsv = testCsv("testCSV")
         private val csvWithFrenchLocale = testCsv("testCSVwithFrenchLocale")

plugins/dataframe-gradle-plugin/src/test/kotlin/org/jetbrains/dataframe/gradle/DataFrameReadTest.kt

@@ -7,6 +7,7 @@ import io.kotest.assertions.throwables.shouldThrowAny
 import io.kotest.matchers.shouldBe
 import kotlinx.serialization.SerializationException
 import org.jetbrains.kotlinx.dataframe.DataFrame
+import org.jetbrains.kotlinx.dataframe.api.isEmpty
 import org.jetbrains.kotlinx.dataframe.io.read
 import org.jetbrains.kotlinx.dataframe.io.readSqlTable
 import org.junit.Test
@@ -42,9 +43,7 @@ class DataFrameReadTest {
     fun `file with invalid csv`() {
         val temp = Files.createTempDirectory("").toFile()
         val invalidCsv = File(temp, "test.csv").also { it.writeText("") }
-        shouldThrow<IndexOutOfBoundsException> {
-            DataFrame.read(invalidCsv)
-        }
+        DataFrame.read(invalidCsv).isEmpty() shouldBe true
     }
 
     @Test