[SPARK-25945][SQL] Support locale while parsing date/timestamp from CSV/JSON

python/pyspark/sql/readwriter.py

@@ -177,7 +177,7 @@ def json(self, path, schema=None, primitivesAsString=None, prefersDecimal=None,
              allowNumericLeadingZero=None, allowBackslashEscapingAnyCharacter=None,
              mode=None, columnNameOfCorruptRecord=None, dateFormat=None, timestampFormat=None,
              multiLine=None, allowUnquotedControlChars=None, lineSep=None, samplingRatio=None,
-             dropFieldIfAllNull=None, encoding=None):
+             dropFieldIfAllNull=None, encoding=None, locale=None):
         """
         Loads JSON files and returns the results as a :class:`DataFrame`.
 
@@ -249,6 +249,9 @@ def json(self, path, schema=None, primitivesAsString=None, prefersDecimal=None,
         :param dropFieldIfAllNull: whether to ignore column of all null values or empty
                                    array/struct during schema inference. If None is set, it
                                    uses the default value, ``false``.
+        :param locale: sets a locale as language tag in IETF BCP 47 format. If None is set,
+                       it uses the default value, ``en-US``. For instance, ``locale`` is used while
+                       parsing dates and timestamps.
 
         >>> df1 = spark.read.json('python/test_support/sql/people.json')
         >>> df1.dtypes
@@ -267,7 +270,8 @@ def json(self, path, schema=None, primitivesAsString=None, prefersDecimal=None,
             mode=mode, columnNameOfCorruptRecord=columnNameOfCorruptRecord, dateFormat=dateFormat,
             timestampFormat=timestampFormat, multiLine=multiLine,
             allowUnquotedControlChars=allowUnquotedControlChars, lineSep=lineSep,
-            samplingRatio=samplingRatio, dropFieldIfAllNull=dropFieldIfAllNull, encoding=encoding)
+            samplingRatio=samplingRatio, dropFieldIfAllNull=dropFieldIfAllNull, encoding=encoding,
+            locale=locale)
         if isinstance(path, basestring):
             path = [path]
         if type(path) == list:
@@ -349,7 +353,7 @@ def csv(self, path, schema=None, sep=None, encoding=None, quote=None, escape=Non
             negativeInf=None, dateFormat=None, timestampFormat=None, maxColumns=None,
             maxCharsPerColumn=None, maxMalformedLogPerPartition=None, mode=None,
             columnNameOfCorruptRecord=None, multiLine=None, charToEscapeQuoteEscaping=None,
-            samplingRatio=None, enforceSchema=None, emptyValue=None):
+            samplingRatio=None, enforceSchema=None, emptyValue=None, locale=None):
         r"""Loads a CSV file and returns the result as a  :class:`DataFrame`.
 
         This function will go through the input once to determine the input schema if
@@ -446,6 +450,9 @@ def csv(self, path, schema=None, sep=None, encoding=None, quote=None, escape=Non
                               If None is set, it uses the default value, ``1.0``.
         :param emptyValue: sets the string representation of an empty value. If None is set, it uses
                            the default value, empty string.
+        :param locale: sets a locale as language tag in IETF BCP 47 format. If None is set,
+                       it uses the default value, ``en-US``. For instance, ``locale`` is used while
+                       parsing dates and timestamps.
 
         >>> df = spark.read.csv('python/test_support/sql/ages.csv')
         >>> df.dtypes
@@ -465,7 +472,7 @@ def csv(self, path, schema=None, sep=None, encoding=None, quote=None, escape=Non
             maxMalformedLogPerPartition=maxMalformedLogPerPartition, mode=mode,
             columnNameOfCorruptRecord=columnNameOfCorruptRecord, multiLine=multiLine,
             charToEscapeQuoteEscaping=charToEscapeQuoteEscaping, samplingRatio=samplingRatio,
-            enforceSchema=enforceSchema, emptyValue=emptyValue)
+            enforceSchema=enforceSchema, emptyValue=emptyValue, locale=locale)
         if isinstance(path, basestring):
             path = [path]
         if type(path) == list:

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/csv/CSVOptions.scala

@@ -131,13 +131,16 @@ class CSVOptions(
   val timeZone: TimeZone = DateTimeUtils.getTimeZone(
     parameters.getOrElse(DateTimeUtils.TIMEZONE_OPTION, defaultTimeZoneId))
 
+  // A language tag in IETF BCP 47 format
+  val locale: Locale = parameters.get("locale").map(Locale.forLanguageTag).getOrElse(Locale.US)
+
   // Uses `FastDateFormat` which can be direct replacement for `SimpleDateFormat` and thread-safe.
   val dateFormat: FastDateFormat =
-    FastDateFormat.getInstance(parameters.getOrElse("dateFormat", "yyyy-MM-dd"), Locale.US)
+    FastDateFormat.getInstance(parameters.getOrElse("dateFormat", "yyyy-MM-dd"), locale)
 
   val timestampFormat: FastDateFormat =
     FastDateFormat.getInstance(
-      parameters.getOrElse("timestampFormat", "yyyy-MM-dd'T'HH:mm:ss.SSSXXX"), timeZone, Locale.US)
+      parameters.getOrElse("timestampFormat", "yyyy-MM-dd'T'HH:mm:ss.SSSXXX"), timeZone, locale)
 
   val multiLine = parameters.get("multiLine").map(_.toBoolean).getOrElse(false)
 

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/json/JSONOptions.scala

@@ -76,16 +76,19 @@ private[sql] class JSONOptions(
   // Whether to ignore column of all null values or empty array/struct during schema inference
   val dropFieldIfAllNull = parameters.get("dropFieldIfAllNull").map(_.toBoolean).getOrElse(false)
 
+  // A language tag in IETF BCP 47 format
+  val locale: Locale = parameters.get("locale").map(Locale.forLanguageTag).getOrElse(Locale.US)
+
   val timeZone: TimeZone = DateTimeUtils.getTimeZone(
     parameters.getOrElse(DateTimeUtils.TIMEZONE_OPTION, defaultTimeZoneId))
 
   // Uses `FastDateFormat` which can be direct replacement for `SimpleDateFormat` and thread-safe.
   val dateFormat: FastDateFormat =
-    FastDateFormat.getInstance(parameters.getOrElse("dateFormat", "yyyy-MM-dd"), Locale.US)
+    FastDateFormat.getInstance(parameters.getOrElse("dateFormat", "yyyy-MM-dd"), locale)
 
   val timestampFormat: FastDateFormat =
     FastDateFormat.getInstance(
-      parameters.getOrElse("timestampFormat", "yyyy-MM-dd'T'HH:mm:ss.SSSXXX"), timeZone, Locale.US)
+      parameters.getOrElse("timestampFormat", "yyyy-MM-dd'T'HH:mm:ss.SSSXXX"), timeZone, locale)
 
   val multiLine = parameters.get("multiLine").map(_.toBoolean).getOrElse(false)
 

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/expressions/CsvExpressionsSuite.scala

@@ -17,7 +17,8 @@
 
 package org.apache.spark.sql.catalyst.expressions
 
-import java.util.Calendar
+import java.text.SimpleDateFormat
+import java.util.{Calendar, Locale}
 
 import org.scalatest.exceptions.TestFailedException
 
@@ -209,4 +210,20 @@ class CsvExpressionsSuite extends SparkFunSuite with ExpressionEvalHelper with P
       "2015-12-31T16:00:00"
     )
   }
+
+  test("parse date with locale") {
+    Seq("en-US", "ru-RU").foreach { langTag =>
+      val locale = Locale.forLanguageTag(langTag)
+      val date = new SimpleDateFormat("yyyy-MM-dd").parse("2018-11-05")
+      val schema = new StructType().add("d", DateType)
+      val dateFormat = "MMM yyyy"
+      val sdf = new SimpleDateFormat(dateFormat, locale)
+      val dateStr = sdf.format(date)
+      val options = Map("dateFormat" -> dateFormat, "locale" -> langTag)
+
+      checkEvaluation(
+        CsvToStructs(schema, options, Literal.create(dateStr), gmtId),
+        InternalRow(17836)) // number of days from 1970-01-01
+    }
+  }
 }

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/expressions/JsonExpressionsSuite.scala

@@ -17,7 +17,8 @@
 
 package org.apache.spark.sql.catalyst.expressions
 
-import java.util.Calendar
+import java.text.SimpleDateFormat
+import java.util.{Calendar, Locale}
 
 import org.scalatest.exceptions.TestFailedException
 
@@ -737,4 +738,20 @@ class JsonExpressionsSuite extends SparkFunSuite with ExpressionEvalHelper with
         CreateMap(Seq(Literal.create("allowNumericLeadingZeros"), Literal.create("true")))),
       "struct<col:bigint>")
   }
+
+  test("parse date with locale") {
+    Seq("en-US", "ru-RU").foreach { langTag =>
+      val locale = Locale.forLanguageTag(langTag)
+      val date = new SimpleDateFormat("yyyy-MM-dd").parse("2018-11-05")
+      val schema = new StructType().add("d", DateType)
+      val dateFormat = "MMM yyyy"
+      val sdf = new SimpleDateFormat(dateFormat, locale)
+      val dateStr = s"""{"d":"${sdf.format(date)}"}"""
+      val options = Map("dateFormat" -> dateFormat, "locale" -> langTag)
+
+      checkEvaluation(
+        JsonToStructs(schema, options, Literal.create(dateStr), gmtId),
+        InternalRow(17836)) // number of days from 1970-01-01
+    }
+  }
 }

sql/core/src/main/scala/org/apache/spark/sql/DataFrameReader.scala

@@ -384,6 +384,8 @@ class DataFrameReader private[sql](sparkSession: SparkSession) extends Logging {
    * for schema inferring.</li>
    * <li>`dropFieldIfAllNull` (default `false`): whether to ignore column of all null values or
    * empty array/struct during schema inference.</li>
+   * <li>`locale` (default is `en-US`): sets a locale as language tag in IETF BCP 47 format.
+   * For instance, this is used while parsing dates and timestamps.</li>
    * </ul>
    *
    * @since 2.0.0
@@ -604,6 +606,8 @@ class DataFrameReader private[sql](sparkSession: SparkSession) extends Logging {
    * `spark.sql.columnNameOfCorruptRecord`): allows renaming the new field having malformed string
    * created by `PERMISSIVE` mode. This overrides `spark.sql.columnNameOfCorruptRecord`.</li>
    * <li>`multiLine` (default `false`): parse one record, which may span multiple lines.</li>
+   * <li>`locale` (default is `en-US`): sets a locale as language tag in IETF BCP 47 format.
+   * For instance, this is used while parsing dates and timestamps.</li>
    * </ul>
    *
    * @since 2.0.0

sql/core/src/main/scala/org/apache/spark/sql/streaming/DataStreamReader.scala

@@ -296,6 +296,8 @@ final class DataStreamReader private[sql](sparkSession: SparkSession) extends Lo
    * that should be used for parsing.</li>
    * <li>`dropFieldIfAllNull` (default `false`): whether to ignore column of all null values or
    * empty array/struct during schema inference.</li>
+   * <li>`locale` (default is `en-US`): sets a locale as language tag in IETF BCP 47 format.
+   * For instance, this is used while parsing dates and timestamps.</li>
    * </ul>
    *
    * @since 2.0.0
@@ -372,6 +374,8 @@ final class DataStreamReader private[sql](sparkSession: SparkSession) extends Lo
    * `spark.sql.columnNameOfCorruptRecord`): allows renaming the new field having malformed string
    * created by `PERMISSIVE` mode. This overrides `spark.sql.columnNameOfCorruptRecord`.</li>
    * <li>`multiLine` (default `false`): parse one record, which may span multiple lines.</li>
+   * <li>`locale` (default is `en-US`): sets a locale as language tag in IETF BCP 47 format.
+   * For instance, this is used while parsing dates and timestamps.</li>
    * </ul>
    *
    * @since 2.0.0

sql/core/src/test/scala/org/apache/spark/sql/CsvFunctionsSuite.scala

@@ -17,6 +17,9 @@
 
 package org.apache.spark.sql
 
+import java.text.SimpleDateFormat
+import java.util.Locale
+
 import scala.collection.JavaConverters._
 
 import org.apache.spark.SparkException
@@ -117,4 +120,18 @@ class CsvFunctionsSuite extends QueryTest with SharedSQLContext {
           "Acceptable modes are PERMISSIVE and FAILFAST."))
     }
   }
+
+  test("parse timestamps with locale") {
+    Seq("en-US", "ko-KR", "zh-CN", "ru-RU").foreach { langTag =>
+      val locale = Locale.forLanguageTag(langTag)
+      val ts = new SimpleDateFormat("dd/MM/yyyy HH:mm").parse("06/11/2018 18:00")
+      val timestampFormat = "dd MMM yyyy HH:mm"
+      val sdf = new SimpleDateFormat(timestampFormat, locale)
+      val input = Seq(s"""${sdf.format(ts)}""").toDS()
+      val options = Map("timestampFormat" -> timestampFormat, "locale" -> langTag)
+      val df = input.select(from_csv($"value", lit("time timestamp"), options.asJava))
+
+      checkAnswer(df, Row(Row(java.sql.Timestamp.valueOf("2018-11-06 18:00:00.0"))))
+    }
+  }
 }

sql/core/src/test/scala/org/apache/spark/sql/JsonFunctionsSuite.scala

@@ -17,6 +17,9 @@
 
 package org.apache.spark.sql
 
+import java.text.SimpleDateFormat
+import java.util.Locale
+
 import collection.JavaConverters._
 
 import org.apache.spark.SparkException
@@ -578,4 +581,18 @@ class JsonFunctionsSuite extends QueryTest with SharedSQLContext {
           "Acceptable modes are PERMISSIVE and FAILFAST."))
     }
   }
+
+  test("parse timestamps with locale") {
+    Seq("en-US", "ko-KR", "zh-CN", "ru-RU").foreach { langTag =>
+      val locale = Locale.forLanguageTag(langTag)
+      val ts = new SimpleDateFormat("dd/MM/yyyy HH:mm").parse("06/11/2018 18:00")
+      val timestampFormat = "dd MMM yyyy HH:mm"
+      val sdf = new SimpleDateFormat(timestampFormat, locale)
+      val input = Seq(s"""{"time": "${sdf.format(ts)}"}""").toDS()
+      val options = Map("timestampFormat" -> timestampFormat, "locale" -> langTag)
+      val df = input.select(from_json($"value", "time timestamp", options))
+
+      checkAnswer(df, Row(Row(java.sql.Timestamp.valueOf("2018-11-06 18:00:00.0"))))
+    }
+  }
 }