[CALCITE-5543] WIP: Implement BigQuery functions for parsing DATE, TIME, TIMESTAMP, DATETIME (#27)

Author: tjbanghart

core/src/main/java/org/apache/calcite/sql/fun/SqlLibraryOperators.java

@@ -881,6 +881,49 @@ static RelDataType deriveTypeSplit(SqlOperatorBinding operatorBinding,
           OperandTypes.STRING_STRING,
           SqlFunctionCategory.TIMEDATE);
 
+  /**
+   * The "PARSE_TIME(string, string)" function (BigQuery); Converts a string representation of time
+   * to a TIME object.
+   */
+  @LibraryOperator(libraries = {BIG_QUERY})
+  public static final SqlFunction PARSE_TIME =
+      SqlBasicFunction.create("PARSE_TIME", ReturnTypes.TIME_NULLABLE,
+          OperandTypes.STRING_STRING, SqlFunctionCategory.TIMEDATE);
+
+  /**
+   * The "PARSE_DATE(string, string)" function (BigQuery); Converts a string representation of date
+   * to a DATE object.
+   */
+  @LibraryOperator(libraries = {BIG_QUERY})
+  public static final SqlFunction PARSE_DATE =
+      SqlBasicFunction.create("PARSE_DATE",
+          ReturnTypes.DATE_NULLABLE, OperandTypes.STRING_STRING, SqlFunctionCategory.TIMEDATE);
+
+  /**
+   * The "PARSE_TIMESTAMP(string, string [, timezone])" function (BigQuery); Formats a timestamp
+   * object according to the specified string.
+   *
+   * <p>In BigQuery, the "TIMESTAMP" datatype maps to Calcite's
+   * TIMESTAMP_WITH_LOCAL_TIME_ZONE
+   */
+  @LibraryOperator(libraries = {BIG_QUERY})
+  public static final SqlFunction PARSE_TIMESTAMP =
+      SqlBasicFunction.create("PARSE_TIMESTAMP",
+          ReturnTypes.TIMESTAMP_LTZ_NULLABLE, OperandTypes.STRING_STRING_OPTIONAL_STRING,
+          SqlFunctionCategory.TIMEDATE);
+
+  /**
+   * The "PARSE_DATETIME(string, string [, timezone])" function (BigQuery); Formats a timestamp
+   * object according to the specified string.
+   *
+   * <p>Note that the {@code TIMESTAMP} type of Calcite and Standard SQL
+   * is called {@code DATETIME} in BigQuery.
+   */
+  @LibraryOperator(libraries = {BIG_QUERY})
+  public static final SqlFunction PARSE_DATETIME =
+      SqlBasicFunction.create("PARSE_DATETIME", ReturnTypes.TIMESTAMP_NULLABLE,
+          OperandTypes.STRING_STRING, SqlFunctionCategory.TIMEDATE);
+
   /** The "FORMAT_TIME(string, time)" function (BigQuery);
    * Formats a time object according to the specified string. */
   @LibraryOperator(libraries = {BIG_QUERY})

core/src/main/java/org/apache/calcite/sql/type/ReturnTypes.java

@@ -352,6 +352,13 @@ public static SqlCall stripSeparator(SqlCall call) {
   public static final SqlReturnTypeInference TIMESTAMP_LTZ =
       explicit(SqlTypeName.TIMESTAMP_WITH_LOCAL_TIME_ZONE);
 
+  /**
+   * Type-inference strategy whereby the result type of a call is nullable
+   * TIMESTAMP WITH LOCAL TIME ZONE.
+   */
+  public static final SqlReturnTypeInference TIMESTAMP_LTZ_NULLABLE =
+      TIMESTAMP_LTZ.andThen(SqlTypeTransforms.TO_NULLABLE);
+
   /**
    * Type-inference strategy whereby the result type of a call is Double.
    */

core/src/test/java/org/apache/calcite/rel/rel2sql/RelToSqlConverterTest.java

@@ -1828,6 +1828,38 @@ private SqlDialect nonOrdinalDialect() {
     sql(query).withBigQuery().ok(expected);
   }
 
+  @Test void testBigQueryParseDatetimeFunctions() {
+    String parseTime = "select parse_time('%I:%M:%S', '07:30:00')\n"
+        + "from \"foodmart\".\"product\"\n";
+    final String expectedTimestampTrunc =
+        "SELECT PARSE_TIME('%I:%M:%S', '07:30:00')\n"
+            + "FROM \"foodmart\".\"product\"";
+    sql(parseTime).withLibrary(SqlLibrary.BIG_QUERY).ok(expectedTimestampTrunc);
+
+    String parseDate = "select parse_date('%A %b %e %Y', 'Thursday Dec 25 2008')\n"
+        + "from \"foodmart\".\"product\"\n";
+    final String expectedParseDate =
+        "SELECT PARSE_DATE('%A %b %e %Y', 'Thursday Dec 25 2008')\n"
+            + "FROM \"foodmart\".\"product\"";
+    sql(parseDate).withLibrary(SqlLibrary.BIG_QUERY).ok(expectedParseDate);
+
+    String parseTimestamp =
+        "select parse_timestamp('%a %b %e %I:%M:%S %Y', 'Thu Dec 25 07:30:00 2008')\n"
+        + "from \"foodmart\".\"product\"\n";
+    final String expectedParseTimestamp =
+        "SELECT PARSE_TIMESTAMP('%a %b %e %I:%M:%S %Y', 'Thu Dec 25 07:30:00 2008')\n"
+            + "FROM \"foodmart\".\"product\"";
+    sql(parseTimestamp).withLibrary(SqlLibrary.BIG_QUERY).ok(expectedParseTimestamp);
+
+    String parseDatetime =
+        "select parse_datetime('%a %b %e %I:%M:%S %Y', 'Thu Dec 25 07:30:00 2008')\n"
+        + "from \"foodmart\".\"product\"\n";
+    final String expectedParseDatetime =
+        "SELECT PARSE_DATETIME('%a %b %e %I:%M:%S %Y', 'Thu Dec 25 07:30:00 2008')\n"
+            + "FROM \"foodmart\".\"product\"";
+    sql(parseDatetime).withLibrary(SqlLibrary.BIG_QUERY).ok(expectedParseDatetime);
+  }
+
   @Test void testBigQueryTimeTruncFunctions() {
     String timestampTrunc = "select timestamp_trunc(timestamp '2012-02-03 15:30:00', month)\n"
         + "from \"foodmart\".\"product\"\n";

site/_docs/reference.md

@@ -2713,6 +2713,10 @@ BigQuery's type system uses confusingly different names for types and functions:
 | b m p | MD5(string)                                | Calculates an MD5 128-bit checksum of *string* and returns it as a hex string
 | m | MONTHNAME(date)                                | Returns the name, in the connection's locale, of the month in *datetime*; for example, it returns '二月' for both DATE '2020-02-10' and TIMESTAMP '2020-02-10 10:10:10'
 | o | NVL(value1, value2)                            | Returns *value1* if *value1* is not null, otherwise *value2*
+| b | PARSE_DATE(string1, string2)                   | Uses format specified by *string1* to convert *string2* representation of date to a DATE object
+| b | PARSE_DATETIME(string1, string2)               | Uses format specified by *string1* to convert *string2* representation of datetime to a TIMESTAMP object
+| b | PARSE_TIME(string1, string2)                   | Uses format specified by *string1* to convert *string2* representation of time to a TIME object
+| b | PARSE_TIMESTAMP(string1, string2[, timeZone])  | Uses format specified by *string1* to convert *string2* representation of timestamp to a TIMESTAMP WITH LOCAL TIME ZONE object in *timeZone*
 | b | POW(numeric1, numeric2)                        | Returns *numeric1* raised to the power *numeric2*
 | m o | REGEXP_REPLACE(string, regexp, rep, [, pos [, occurrence [, matchType]]]) | Replaces all substrings of *string* that match *regexp* with *rep* at the starting *pos* in expr (if omitted, the default is 1), *occurrence* means which occurrence of a match to search for (if omitted, the default is 1), *matchType* specifies how to perform matching
 | b m p | REPEAT(string, integer)                    | Returns a string consisting of *string* repeated of *integer* times; returns an empty string if *integer* is less than 1