chore: Improve error handling

Signed-off-by: Alex Qyoun-ae <[email protected]>

Author: MazterQyou

datafusion/common/src/error.rs

@@ -153,8 +153,7 @@ impl Display for DataFusionError {
                 write!(f, "This feature is not implemented: {}", desc)
             }
             DataFusionError::Internal(ref desc) => {
-                write!(f, "Internal error: {}. This was likely caused by a bug in DataFusion's \
-                    code and we would welcome that you file an bug report in our issue tracker", desc)
+                write!(f, "Internal error: {}", desc)
             }
             DataFusionError::Plan(ref desc) => {
                 write!(f, "Error during planning: {}", desc)

datafusion/core/src/physical_plan/repartition.rs

@@ -27,7 +27,10 @@ use crate::error::{DataFusionError, Result};
 use crate::physical_plan::hash_utils::create_hashes;
 use crate::physical_plan::{DisplayFormatType, ExecutionPlan, Partitioning, Statistics};
 use arrow::record_batch::RecordBatch;
-use arrow::{array::Array, error::Result as ArrowResult};
+use arrow::{
+    array::Array,
+    error::{ArrowError, Result as ArrowResult},
+};
 use arrow::{compute::take, datatypes::SchemaRef};
 use tokio_stream::wrappers::UnboundedReceiverStream;
 
@@ -369,9 +372,8 @@ impl RepartitionExec {
                             .columns()
                             .iter()
                             .map(|c| {
-                                take(c.as_ref(), &indices, None).map_err(|e| {
-                                    DataFusionError::Execution(e.to_string())
-                                })
+                                take(c.as_ref(), &indices, None)
+                                    .map_err(DataFusionError::ArrowError)
                             })
                             .collect::<Result<Vec<Arc<dyn Array>>>>()?;
                         let output_batch =
@@ -426,9 +428,31 @@ impl RepartitionExec {
             }
             // Error from running input task
             Ok(Err(e)) => {
+                // try to unwrap nested errors
+                let mut err = &e;
+                let message;
+                loop {
+                    if let DataFusionError::External(ext_err) = err {
+                        message = ext_err.to_string();
+                        break;
+                    }
+                    let DataFusionError::ArrowError(arrow_err) = err else {
+                        message = err.to_string();
+                        break;
+                    };
+                    let ArrowError::ExternalError(ext_err) = arrow_err else {
+                        message = arrow_err.to_string();
+                        break;
+                    };
+                    let Some(df_err) = ext_err.downcast_ref::<DataFusionError>() else {
+                        message = ext_err.to_string();
+                        break;
+                    };
+                    err = df_err;
+                }
                 for (_, tx) in txs {
                     // wrap it because need to send error to all output partitions
-                    let err = DataFusionError::Execution(e.to_string());
+                    let err = DataFusionError::Execution(message.clone());
                     let err = Err(err.into());
                     tx.send(Some(err)).ok();
                 }