working on review feedback

Author: Jolanrensen

dataframe-csv/src/main/kotlin/org/jetbrains/kotlinx/dataframe/impl/io/readDelim.kt

@@ -249,6 +249,23 @@ private fun CsvSpecs.Builder.skipLines(takeHeaderFromCsv: Boolean, skipLines: Lo
 /**
  * Sets the correct parsers for the csv, based on [colTypes] and [ParserOptions.skipTypes].
  * If [ColType.DEFAULT] is present, it sets the default parser.
+ *
+ * Logic overview:
+ *
+ * - if no [colTypes] are given
+ *     - let deephaven use all its [default parsers][Parsers.DEFAULT]
+ *     - subtract parsers of [skipTypes][ParserOptions.skipTypes] if those are supplied
+ * - if [colTypes] are supplied
+ *     - if [ColType.DEFAULT] is among the values
+ *       - set the parser for each supplied column+colType
+ *       - let deephaven use _only_ the parser given as [ColType.DEFAULT] type
+ *     - if [ColType.DEFAULT] is not among the values
+ *       - set the parser for each supplied column+coltype
+ *       - let deephaven use all its [default parsers][Parsers.DEFAULT]
+ *       - subtract parsers of [skipTypes][ParserOptions.skipTypes] if those are supplied
+ *
+ * Note that `skipTypes` will never skip a type explicitly set by `colTypes`.
+ * This is intended.
  */
 private fun CsvSpecs.Builder.parsers(
     parserOptions: ParserOptions,

dataframe-csv/src/test/kotlin/org/jetbrains/kotlinx/dataframe/io/DelimCsvTsvTests.kt

@@ -433,6 +433,39 @@ class DelimCsvTsvTests {
         shouldThrow<IllegalArgumentException> { DataFrame.readCsv("NON EXISTENT FILE") }
     }
 
+    @Test
+    fun `cannot auto-parse specific date string`() {
+        @Language("csv")
+        val frenchCsv =
+            """
+            name; price; date;
+            a;12,45; 05/06/2021;
+            b;-13,35;14/07/2025;
+            c;100 123,35;;
+            d;-204 235,23;;
+            e;NaN;;
+            f;null;;
+            """.trimIndent()
+
+        val dfDeephaven = DataFrame.readCsvStr(
+            text = frenchCsv,
+            delimiter = ';',
+        )
+
+        // could not parse, remains String
+        dfDeephaven["date"].type() shouldBe typeOf<String?>()
+
+        val dfDataFrame = DataFrame.readCsvStr(
+            text = frenchCsv,
+            delimiter = ';',
+            // setting any locale skips deephaven's date parsing
+            parserOptions = DEFAULT_PARSER_OPTIONS.copy(locale = Locale.ROOT),
+        )
+
+        // could not parse, remains String
+        dfDataFrame["date"].type() shouldBe typeOf<String?>()
+    }
+
     @Test
     fun `parse with other locales`() {
         @Language("csv")
@@ -626,12 +659,13 @@ class DelimCsvTsvTests {
     fun `skipping types`() {
         val irisDataset = DataFrame.readCsv(
             irisDataset,
+            colTypes = mapOf("sepal.length" to ColType.Double),
             parserOptions = DEFAULT_PARSER_OPTIONS.copy(
                 skipTypes = setOf(typeOf<Double>()),
             ),
         )
 
-        irisDataset["sepal.length"].type() shouldBe typeOf<BigDecimal>()
+        irisDataset["sepal.length"].type() shouldBe typeOf<Double>()
         irisDataset["sepal.width"].type() shouldBe typeOf<BigDecimal>()
         irisDataset["petal.length"].type() shouldBe typeOf<BigDecimal>()
         irisDataset["petal.width"].type() shouldBe typeOf<BigDecimal>()