largeutf to temporal (#8357)

Signed-off-by: jayzhan211 <[email protected]>
apache · Nov 30, 2023 · a588123 · a588123
1 parent c079a92
commit a588123
Show file tree

Hide file tree

Showing 2 changed files with 36 additions and 18 deletions.
diff --git a/datafusion/expr/src/type_coercion/binary.rs b/datafusion/expr/src/type_coercion/binary.rs
@@ -331,26 +331,27 @@ fn string_temporal_coercion(
     rhs_type: &DataType,
 ) -> Option<DataType> {
     use arrow::datatypes::DataType::*;
-    match (lhs_type, rhs_type) {
-        (Utf8, Date32) | (Date32, Utf8) => Some(Date32),
-        (Utf8, Date64) | (Date64, Utf8) => Some(Date64),
-        (Utf8, Time32(unit)) | (Time32(unit), Utf8) => {
-            match is_time_with_valid_unit(Time32(unit.clone())) {
-                false => None,
-                true => Some(Time32(unit.clone())),
-            }
-        }
-        (Utf8, Time64(unit)) | (Time64(unit), Utf8) => {
-            match is_time_with_valid_unit(Time64(unit.clone())) {
-                false => None,
-                true => Some(Time64(unit.clone())),
-            }
-        }
-        (Timestamp(_, tz), Utf8) | (Utf8, Timestamp(_, tz)) => {
-            Some(Timestamp(TimeUnit::Nanosecond, tz.clone()))
+
+    fn match_rule(l: &DataType, r: &DataType) -> Option<DataType> {
+        match (l, r) {
+            // Coerce Utf8/LargeUtf8 to Date32/Date64/Time32/Time64/Timestamp
+            (Utf8, temporal) | (LargeUtf8, temporal) => match temporal {
+                Date32 | Date64 => Some(temporal.clone()),
+                Time32(_) | Time64(_) => {
+                    if is_time_with_valid_unit(temporal.to_owned()) {
+                        Some(temporal.to_owned())
+                    } else {
+                        None
+                    }
+                }
+                Timestamp(_, tz) => Some(Timestamp(TimeUnit::Nanosecond, tz.clone())),
+                _ => None,
+            },
+            _ => None,
         }
-        _ => None,
     }
+
+    match_rule(lhs_type, rhs_type).or_else(|| match_rule(rhs_type, lhs_type))
 }
 
 /// Coerce `lhs_type` and `rhs_type` to a common type for the purposes of a comparison operation

diff --git a/datafusion/sqllogictest/test_files/scalar.slt b/datafusion/sqllogictest/test_files/scalar.slt
@@ -1926,3 +1926,20 @@ A true
 B false
 C false
 D false
+
+# test string_temporal_coercion
+query BBBBBBBBBB
+select 
+  arrow_cast(to_timestamp('2020-01-01 01:01:11.1234567890Z'), 'Timestamp(Second, None)') == '2020-01-01T01:01:11',
+  arrow_cast(to_timestamp('2020-01-02 01:01:11.1234567890Z'), 'Timestamp(Second, None)') == arrow_cast('2020-01-02T01:01:11', 'LargeUtf8'),
+  arrow_cast(to_timestamp('2020-01-03 01:01:11.1234567890Z'), 'Time32(Second)') == '01:01:11',
+  arrow_cast(to_timestamp('2020-01-04 01:01:11.1234567890Z'), 'Time32(Second)') == arrow_cast('01:01:11', 'LargeUtf8'),
+  arrow_cast(to_timestamp('2020-01-05 01:01:11.1234567890Z'), 'Time64(Microsecond)') == '01:01:11.123456',
+  arrow_cast(to_timestamp('2020-01-06 01:01:11.1234567890Z'), 'Time64(Microsecond)') == arrow_cast('01:01:11.123456', 'LargeUtf8'),
+  arrow_cast('2020-01-07', 'Date32') == '2020-01-07',
+  arrow_cast('2020-01-08', 'Date64') == '2020-01-08',
+  arrow_cast('2020-01-09', 'Date32') == arrow_cast('2020-01-09', 'LargeUtf8'),
+  arrow_cast('2020-01-10', 'Date64') == arrow_cast('2020-01-10', 'LargeUtf8')
+;
+----
+true true true true true true true true true true