[SPARK-19488][SQL] fix csv infer schema when the field is Nan/Inf etc

## What changes were proposed in this pull request?

when csv infer schema, it does not use user defined csvoptions to parse the field, such as `inf`, `-inf` which are should be parsed to DoubleType

this pr add  `options.nanValue`, `options.negativeInf`, `options.positiveIn`  to check if the field is a DoubleType

## How was this patch tested?
unit test added

Author: windpiger <songjun@outlook.com>

Closes #16834 from windpiger/fixinferInfSchemaCsv.

1 files changed, 8 insertions, 0 deletions

diff --git a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVInferSchemaSuite.scala b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVInferSchemaSuite.scala
index 8620bb9f65..d8c6c25504 100644
--- a/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVInferSchemaSuite.scala
+++ b/sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/csv/CSVInferSchemaSuite.scala
@@ -131,4 +131,12 @@ class CSVInferSchemaSuite extends SparkFunSuite {
     assert(CSVInferSchema.inferField(DecimalType(20, 0), "2015-12-01 00:00:00", options)
       == StringType)
   }
+
+  test("DoubleType should be infered when user defined nan/inf are provided") {
+    val options = new CSVOptions(Map("nanValue" -> "nan", "negativeInf" -> "-inf",
+      "positiveInf" -> "inf"))
+    assert(CSVInferSchema.inferField(NullType, "nan", options) == DoubleType)
+    assert(CSVInferSchema.inferField(NullType, "inf", options) == DoubleType)
+    assert(CSVInferSchema.inferField(NullType, "-inf", options) == DoubleType)
+  }
 }