[Feature][Transforms] Support is_date function of sql (#6966)

apache · Jun 12, 2024 · 96fdade · 96fdade
1 parent d56d64f
commit 96fdade
Show file tree

Hide file tree

Showing 5 changed files with 32 additions and 2 deletions.
diff --git a/docs/en/transform-v2/sql-functions.md b/docs/en/transform-v2/sql-functions.md
@@ -795,6 +795,15 @@ Example:
 
 MONTHNAME(CREATED)
 
+### IS_DATE
+
+```IS_DATE(string, formatString)```
+Parses a string and returns a boolean value. The most important format characters are: y year, M month, d day, H hour, m minute, s second. For details of the format, see java.time.format.DateTimeFormatter.
+
+Example:
+
+CALL IS_DATE('2021-04-08 13:34:45','yyyy-MM-dd HH:mm:ss')
+
 ### PARSEDATETIME / TO_DATE
 
 ```PARSEDATETIME | TO_DATE(string, formatString)```

diff --git a/...2e/seatunnel-transforms-v2-e2e-part-2/src/test/resources/sql_transform/func_datetime.conf b/...2e/seatunnel-transforms-v2-e2e-part-2/src/test/resources/sql_transform/func_datetime.conf
@@ -37,10 +37,11 @@ source {
         c4 = "timestamp"
         c5 = "string"
         c6 = "string"
+        c7 = "string"
       }
     }
     rows = [
-      {fields = [1, "Joy Ding", "2021-04-15T13:34:45", "2022-01-23T12:34:56", "2021-04-15T13:34:45.235", "2021-04-08T13:34:45.235", "2021-04-08 13:34:45.235", "2021-04-08"], kind = INSERT}
+      {fields = [1, "Joy Ding", "2021-04-15T13:34:45", "2022-01-23T12:34:56", "2021-04-15T13:34:45.235", "2021-04-08T13:34:45.235", "2021-04-08 13:34:45.235", "2021-04-08", "2021-04-08 13:34:45.235"], kind = INSERT}
     ]
   }
 }
@@ -49,7 +50,7 @@ transform {
   Sql {
     source_table_name = "fake"
     result_table_name = "fake1"
-    query = "select current_date as cd, current_timestamp as ct2, dateadd(c1, 1) as c1_1, dateadd(c1, 40, 'DAY') as c1_2, dateadd(c1, 2, 'YEAR') as c1_3, dateadd(c1, 10, 'MONTH') as c1_4, dateadd(c1, 13, 'HOUR') as c1_5, dateadd(c1, 40, 'MINUTE') as c1_6, dateadd(c1, 30, 'SECOND') as c1_7, datediff(c1, c2) as test, datediff(c1, c2, 'DAY') as c2_1, datediff(c1, c2, 'YEAR') as c2_2, datediff(c1, c2, 'MONTH') as c2_3, datediff(c1, c2, 'HOUR') as c2_4, datediff(c1, c2, 'MINUTE') as c2_5, datediff(c1, c2, 'SECOND') as c2_6, date_trunc(c3, 'YEAR') as c3_1, date_trunc(c3, 'MONTH') as c3_2, date_trunc(c3, 'DAY') as c3_3, date_trunc(c3, 'HOUR') as c3_4, date_trunc(c3, 'MINUTE') as c3_5, date_trunc(c3, 'SECOND') as c3_6, dayname(c3) as c3_7, day_of_week(c3) c3_8, day_of_year(c3) c3_9, extract(YEAR FROM c3) c3_10, extract(MONTH FROM c3) c3_11, extract(DAY FROM c3) c3_12, extract(HOUR FROM c3) c3_13, extract(MINUTE from c3) c3_14, extract(SECOND from c3) c3_15, extract(MILLISECOND from c3) c3_16, extract(DAYOFWEEK FROM c3) c3_17, extract(DAYOFYEAR FROM c3) c3_18, formatdatetime(c4,'yyyy-MM-dd HH:mm:ss.S') c4_1, formatdatetime(c4,'yyyy-MM-dd') c4_2, formatdatetime(c4,'HH:mm:ss.SSS') c4_3, hour(c4) c4_4, minute(c4) c4_5, month(c4) c4_6, monthname(c4) c4_7, parsedatetime(c5,'yyyy-MM-dd HH:mm:ss.SSS') c5_1, to_date(c6,'yyyy-MM-dd') c6_1, quarter(c4) c4_8, second(c4) c4_9, week(c4) c4_10, year(c4) c4_11 from fake"
+    query = "select current_date as cd, current_timestamp as ct2, dateadd(c1, 1) as c1_1, dateadd(c1, 40, 'DAY') as c1_2, dateadd(c1, 2, 'YEAR') as c1_3, dateadd(c1, 10, 'MONTH') as c1_4, dateadd(c1, 13, 'HOUR') as c1_5, dateadd(c1, 40, 'MINUTE') as c1_6, dateadd(c1, 30, 'SECOND') as c1_7, datediff(c1, c2) as test, datediff(c1, c2, 'DAY') as c2_1, datediff(c1, c2, 'YEAR') as c2_2, datediff(c1, c2, 'MONTH') as c2_3, datediff(c1, c2, 'HOUR') as c2_4, datediff(c1, c2, 'MINUTE') as c2_5, datediff(c1, c2, 'SECOND') as c2_6, date_trunc(c3, 'YEAR') as c3_1, date_trunc(c3, 'MONTH') as c3_2, date_trunc(c3, 'DAY') as c3_3, date_trunc(c3, 'HOUR') as c3_4, date_trunc(c3, 'MINUTE') as c3_5, date_trunc(c3, 'SECOND') as c3_6, dayname(c3) as c3_7, day_of_week(c3) c3_8, day_of_year(c3) c3_9, extract(YEAR FROM c3) c3_10, extract(MONTH FROM c3) c3_11, extract(DAY FROM c3) c3_12, extract(HOUR FROM c3) c3_13, extract(MINUTE from c3) c3_14, extract(SECOND from c3) c3_15, extract(MILLISECOND from c3) c3_16, extract(DAYOFWEEK FROM c3) c3_17, extract(DAYOFYEAR FROM c3) c3_18, formatdatetime(c4,'yyyy-MM-dd HH:mm:ss.S') c4_1, formatdatetime(c4,'yyyy-MM-dd') c4_2, formatdatetime(c4,'HH:mm:ss.SSS') c4_3, hour(c4) c4_4, minute(c4) c4_5, month(c4) c4_6, monthname(c4) c4_7, parsedatetime(c5,'yyyy-MM-dd HH:mm:ss.SSS') c5_1, to_date(c6,'yyyy-MM-dd') c6_1, quarter(c4) c4_8, second(c4) c4_9, week(c4) c4_10, year(c4) c4_11, case when c7 is not null and is_date(c7, 'yyyy-MM-dd HH:mm:ss.SSS') then to_date(c7,'yyyy-MM-dd HH:mm:ss.SSS') else null end as c7_1 from fake"
   }
 }
 
@@ -387,6 +388,13 @@ sink {
           field_value = [
             {equals_to = 2021}
           ]
+        },
+        {
+          field_name = "c7_1"
+          field_type = "timestamp"
+          field_value = [
+            {equals_to = "2021-04-08T13:34:45.235"}
+          ]
         }
       ]
     }

diff --git a/...-transforms-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/ZetaSQLFunction.java b/...-transforms-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/ZetaSQLFunction.java
@@ -159,6 +159,7 @@ public class ZetaSQLFunction {
     public static final String MONTHNAME = "MONTHNAME";
     public static final String PARSEDATETIME = "PARSEDATETIME";
     public static final String TO_DATE = "TO_DATE";
+    public static final String IS_DATE = "IS_DATE";
     public static final String QUARTER = "QUARTER";
     public static final String SECOND = "SECOND";
     public static final String WEEK = "WEEK";
@@ -476,6 +477,8 @@ public Object executeFunctionExpr(String functionName, List<Object> args) {
             case PARSEDATETIME:
             case TO_DATE:
                 return DateTimeFunction.parsedatetime(args);
+            case IS_DATE:
+                return DateTimeFunction.isDate(args);
             case QUARTER:
                 return DateTimeFunction.quarter(args);
             case SECOND:

diff --git a/...nnel-transforms-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/ZetaSQLType.java b/...nnel-transforms-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/ZetaSQLType.java
@@ -391,6 +391,7 @@ private SeaTunnelDataType<?> getFunctionType(Function function) {
             case ZetaSQLFunction.DATEDIFF:
                 return BasicType.LONG_TYPE;
             case ZetaSQLFunction.REGEXP_LIKE:
+            case ZetaSQLFunction.IS_DATE:
                 return BasicType.BOOLEAN_TYPE;
             case ZetaSQLFunction.ACOS:
             case ZetaSQLFunction.ASIN:

diff --git a/...-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/functions/DateTimeFunction.java b/...-v2/src/main/java/org/apache/seatunnel/transform/sql/zeta/functions/DateTimeFunction.java
@@ -472,6 +472,15 @@ public static String monthname(List<Object> args) {
         return getMonthsAndWeeks(0)[dow - 1];
     }
 
+    public static boolean isDate(List<Object> args) {
+        try {
+            parsedatetime(args);
+            return true;
+        } catch (Throwable e) {
+            return false;
+        }
+    }
+
     public static Temporal parsedatetime(List<Object> args) {
         String str = (String) args.get(0);
         if (str == null) {