fix: make parse_date able to parse partial dates (#8330)

* fix: make parse_date able to parse partial dates

* use SimpleDateFormat

* use FastDateFormat

* errors

ksqldb-common/pom.xml

@@ -74,7 +74,7 @@
         <dependency>
             <groupId>org.apache.commons</groupId>
             <artifactId>commons-lang3</artifactId>
-            <version>3.3.1</version>
+            <version>3.5</version>
         </dependency>
 
         <dependency>

ksqldb-engine/src/main/java/io/confluent/ksql/function/udf/datetime/ParseDate.java

@@ -25,33 +25,29 @@
 import io.confluent.ksql.function.udf.UdfParameter;
 import io.confluent.ksql.util.KsqlConstants;
 import java.sql.Date;
-import java.time.LocalDate;
-import java.time.format.DateTimeFormatter;
-import java.time.format.DateTimeFormatterBuilder;
-import java.time.temporal.ChronoField;
-import java.time.temporal.TemporalAccessor;
-import java.util.Arrays;
-import java.util.Optional;
+import java.text.ParseException;
+import java.util.TimeZone;
 import java.util.concurrent.ExecutionException;
 import java.util.concurrent.TimeUnit;
+import org.apache.commons.lang3.time.FastDateFormat;
 
 @UdfDescription(
     name = "parse_date",
     category = FunctionCategory.DATE_TIME,
     author = KsqlConstants.CONFLUENT_AUTHOR,
     description = "Converts a string representation of a date in the given format"
         + " into a DATE value. The format pattern should be in the format expected by"
-        + " java.time.format.DateTimeFormatter"
+        + " java.text.SimpleDateFormat"
 )
 public class ParseDate {
 
-  private final LoadingCache<String, DateTimeFormatter> formatters =
+  private static final long MILLIS_IN_DAY = TimeUnit.DAYS.toMillis(1);
+
+  private final LoadingCache<String, FastDateFormat> formatters =
       CacheBuilder.newBuilder()
           .maximumSize(1000)
-          .build(CacheLoader.from(pattern -> new DateTimeFormatterBuilder()
-              .parseCaseInsensitive()
-              .appendPattern(pattern)
-              .toFormatter()));
+          .build(CacheLoader.from(pattern ->
+              FastDateFormat.getInstance(pattern, TimeZone.getTimeZone("GMT"))));
 
   @Udf(description = "Converts a string representation of a date in the given format"
       + " into a DATE value.")
@@ -60,22 +56,14 @@ public Date parseDate(
           description = "The string representation of a date.") final String formattedDate,
       @UdfParameter(
           description = "The format pattern should be in the format expected by"
-              + " java.time.format.DateTimeFormatter.") final String formatPattern) {
+              + " java.text.SimpleDateFormat.") final String formatPattern) {
     try {
-
-      final TemporalAccessor ta = formatters.get(formatPattern).parse(formattedDate);
-      final Optional<ChronoField> timeField = Arrays.stream(ChronoField.values())
-          .filter(field -> field.isTimeBased())
-          .filter(field -> ta.isSupported(field))
-          .findFirst();
-
-      if (timeField.isPresent()) {
+      final long time = formatters.get(formatPattern).parse(formattedDate).getTime();
+      if (time % MILLIS_IN_DAY != 0) {
         throw new KsqlFunctionException("Date format contains time field.");
       }
-
-      return new Date(
-          TimeUnit.DAYS.toMillis(LocalDate.from(ta).toEpochDay()));
-    } catch (final ExecutionException | RuntimeException e) {
+      return new Date(time);
+    } catch (final ExecutionException | RuntimeException | ParseException e) {
       throw new KsqlFunctionException("Failed to parse date '" + formattedDate
           + "' with formatter '" + formatPattern
           + "': " + e.getMessage(), e);

ksqldb-engine/src/test/java/io/confluent/ksql/function/udf/datetime/ParseDateTest.java

@@ -45,6 +45,24 @@ public void shouldConvertStringToDate() {
     assertThat(result.getTime(), is(1638316800000L));
   }
 
+  @Test
+  public void shouldConvertYearMonthToDate() {
+    // When:
+    final Date result = udf.parseDate("2021-12", "yyyy-MM");
+
+    // Then:
+    assertThat(result.getTime(), is(1638316800000L));
+  }
+
+  @Test
+  public void shouldConvertYearToDate() {
+    // When:
+    final Date result = udf.parseDate("2022", "yyyy");
+
+    // Then:
+    assertThat(result.getTime(), is(1640995200000L));
+  }
+
   @Test
   public void shouldConvertCaseInsensitiveStringToDate() {
     // When:

ksqldb-functional-tests/src/test/resources/query-validation-tests/parse-date.json

@@ -11,13 +11,15 @@
         {"topic": "test_topic", "key": "1", "value": "1,zero,11/05/2019,dd/MM/yyyy"},
         {"topic": "test_topic", "key": "2", "value": "2,zero,01-Jan-2022,dd-MMM-yyyy"},
         {"topic": "test_topic", "key": "3", "value": "3,yyy,01-01-1970,dd-MM-yyyy"},
-        {"topic": "test_topic", "key": "4", "value": "4,yyy,01-JAN-2022,dd-MMM-yyyy"}
+        {"topic": "test_topic", "key": "4", "value": "4,yyy,01-JAN-2022,dd-MMM-yyyy"},
+        {"topic": "test_topic", "key": "4", "value": "4,yyy,JAN-2022,MMM-yyyy"}
       ],
       "outputs": [
         {"topic": "TS", "key": "0", "value": "0,17662"},
         {"topic": "TS", "key": "1", "value": "1,18027"},
         {"topic": "TS", "key": "2", "value": "2,18993"},
         {"topic": "TS", "key": "3", "value": "3,0"},
+        {"topic": "TS", "key": "4", "value": "4,18993"},
         {"topic": "TS", "key": "4", "value": "4,18993"}
       ]
     }

pom.xml

@@ -114,7 +114,7 @@
         <commons-text.version>1.8</commons-text.version>
         <csv.version>1.4</csv.version>
         <commons.compress.version>1.21</commons.compress.version>
-        <lang3.version>3.3.1</lang3.version>
+        <lang3.version>3.5</lang3.version>
         <guava.version>30.1.1-jre</guava.version>
         <protobuf.version>3.17.0</protobuf.version>
         <retrying.version>2.0.0</retrying.version>