native_lpad_support

Author: coderfender

native/spark-expr/src/comet_scalar_funcs.rs

@@ -22,9 +22,8 @@ use crate::{
     spark_array_repeat, spark_ceil, spark_date_add, spark_date_sub, spark_decimal_div,
     spark_decimal_integral_div, spark_floor, spark_hex, spark_isnan, spark_lpad,
     spark_make_decimal, spark_read_side_padding, spark_round, spark_rpad, spark_unhex,
-    spark_unscaled_value, SparkBitwiseCount, SparkBitwiseGet, SparkBitwiseNot, SparkDateTrunc,
+    spark_unscaled_value, EvalMode, SparkBitwiseCount, SparkBitwiseNot, SparkDateTrunc,
     SparkStringSpace,
-  
 };
 use arrow::datatypes::DataType;
 use datafusion::common::{DataFusionError, Result as DataFusionResult};

native/spark-expr/src/static_invoke/char_varchar_utils/read_side_padding.rs

@@ -36,6 +36,7 @@ pub fn spark_rpad(args: &[ColumnarValue]) -> Result<ColumnarValue, DataFusionErr
     spark_read_side_padding2(args, true, false)
 }
 
+/// Custom `rpad` because DataFusion's `lpad` has differences in unicode handling
 pub fn spark_lpad(args: &[ColumnarValue]) -> Result<ColumnarValue, DataFusionError> {
     spark_read_side_padding2(args, true, true)
 }
@@ -269,12 +270,14 @@ fn add_padding_string(
             Ok(string)
         }
     } else {
-        let pad_needed = length - char_len;
-        let pad: String = pad_string.chars().cycle().take(pad_needed).collect();
+        let mut result = String::with_capacity(string.len() + pad_string.len());
         if is_left_pad {
-            Ok(format!("{}{}", pad, string))
+            result.push_str(&pad_string);
+            result.push_str(&string);
         } else {
-            Ok(string + &pad)
+            result.push_str(&string);
+            result.push_str(&pad_string);
         }
+        Ok(result)
     }
 }