You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
// Merge (MERGE INTO) is a complex DML operation that requires mutable target state.
1038
-
// DataFusion is primarily a read-only query engine; full MERGE support
1039
-
// needs integration with a table provider that supports writes (e.g. Delta Lake).
1040
-
todo!("merge transform requires a writable table provider (e.g. delta-rs)")
1037
+
Err(TeckelError::Execution(
1038
+
"MERGE transformation requires a mutable table provider. Use the Spark backend for full MERGE INTO support, or decompose into join + filter + union.".to_string()
1039
+
))
1041
1040
}
1042
1041
Source::Parse(t) => {
1043
1042
let df = get(cache,&t.from)?;
@@ -1076,21 +1075,98 @@ pub async fn apply(
1076
1075
}
1077
1076
}
1078
1077
teckel_model::types::ParseFormat::Csv => {
1079
-
// CSV parsing from a string column is complex; use todo for now
1080
-
todo!("parse CSV from string column not yet implemented for DataFusion backend")
1078
+
Err(TeckelError::Execution(
1079
+
"Parse transformation for CSV format is not yet supported in DataFusion. Use the Spark backend or decompose with SQL expressions.".to_string()
"merge not yet implemented for polars backend".to_string(),
549
+
"MERGE transformation requires a mutable table provider. Use the Spark backend for full MERGE INTO support, or decompose into join + filter + union.".to_string(),
550
550
)),
551
551
Source::Parse(_) => Err(TeckelError::Execution(
552
-
"parse not yet implemented for polars backend".to_string(),
0 commit comments