Support Utf8View for string function bit_length() (#13221)

* Support `Utf8View` for string function `bit_length()` Signed-off-by: Austin Liu <[email protected]> * Add scalar test case Signed-off-by: Austin Liu <[email protected]> * Refine tests Signed-off-by: Austin Liu <[email protected]> * Fix wrong format Signed-off-by: Austin Liu <[email protected]> --------- Signed-off-by: Austin Liu <[email protected]>
apache · Nov 10, 2024 · 1557fce · 1557fce
1 parent ee56d6a
commit 1557fce
Show file tree

Hide file tree

Showing 3 changed files with 45 additions and 0 deletions.
diff --git a/datafusion/functions/src/string/bit_length.rs b/datafusion/functions/src/string/bit_length.rs
@@ -79,6 +79,9 @@ impl ScalarUDFImpl for BitLengthFunc {
                 ScalarValue::LargeUtf8(v) => Ok(ColumnarValue::Scalar(
                     ScalarValue::Int64(v.as_ref().map(|x| (x.len() * 8) as i64)),
                 )),
+                ScalarValue::Utf8View(v) => Ok(ColumnarValue::Scalar(
+                    ScalarValue::Int32(v.as_ref().map(|x| (x.len() * 8) as i32)),
+                )),
                 _ => unreachable!("bit length"),
             },
         }

diff --git a/datafusion/sqllogictest/test_files/string/string_literal.slt b/datafusion/sqllogictest/test_files/string/string_literal.slt
@@ -1623,3 +1623,44 @@ a\_c \%abc false
 \%abc a\_c false
 \%abc %abc true
 \%abc \%abc false
+
+# test utf8, largeutf8, utf8view, DictionaryString for bit_length
+query IIII
+SELECT
+  bit_length('Andrew'),
+  bit_length('datafusion数据融合'),
+  bit_length('💖'),
+  bit_length('josé')
+;
+----
+48 176 32 40
+
+query IIII
+SELECT
+  bit_length(arrow_cast('Andrew', 'LargeUtf8')),
+  bit_length(arrow_cast('datafusion数据融合', 'LargeUtf8')),
+  bit_length(arrow_cast('💖', 'LargeUtf8')),
+  bit_length(arrow_cast('josé', 'LargeUtf8'))
+;
+----
+48 176 32 40
+
+query IIII
+SELECT
+  bit_length(arrow_cast('Andrew', 'Utf8View')),
+  bit_length(arrow_cast('datafusion数据融合', 'Utf8View')),
+  bit_length(arrow_cast('💖', 'Utf8View')),
+  bit_length(arrow_cast('josé', 'Utf8View'))
+;
+----
+48 176 32 40
+
+query IIII
+SELECT
+  bit_length(arrow_cast('Andrew', 'Dictionary(Int32, Utf8)')),
+  bit_length(arrow_cast('datafusion数据融合', 'Dictionary(Int32, Utf8)')),
+  bit_length(arrow_cast('💖', 'Dictionary(Int32, Utf8)')),
+  bit_length(arrow_cast('josé', 'Dictionary(Int32, Utf8)'))
+;
+----
+48 176 32 40
diff --git a/datafusion/sqllogictest/test_files/string/string_view.slt b/datafusion/sqllogictest/test_files/string/string_view.slt
@@ -93,6 +93,7 @@ select octet_length(column1_utf8view) from test;
 0
 NULL
 
+# TODO: Revisit this issue after upgrading to the arrow-rs version that includes apache/arrow-rs#6671.
 query error DataFusion error: Arrow error: Compute error: bit_length not supported for Utf8View
 select bit_length(column1_utf8view) from test;