vcftools: minor changes after review

Author: psuszyns

Cargo.toml

@@ -18,3 +18,4 @@ log = "0.4.27"
 serial_test = "3.1"
 tempfile = "3.15"
 parquet = "56.1"
+async-trait = "0.1.88"

datafusion/bio-function-vcftools/Cargo.toml

@@ -13,7 +13,7 @@ datafusion = { workspace = true }
 tokio = { workspace = true }
 futures = { workspace = true }
 log = { workspace = true }
-async-trait = "0.1.88"
+async-trait = { workspace = true }
 
 [dev-dependencies]
 tokio = { workspace = true, features = ["rt-multi-thread", "macros"] }

datafusion/bio-function-vcftools/src/physical/fused_array_transform_exec.rs

@@ -382,7 +382,7 @@ impl FusedArrayTransformStream {
         let mut all_values: Vec<ArrayRef> = Vec::with_capacity(ctx.num_rows);
         let mut offsets: Vec<i32> = vec![0];
         let mut null_bitmap: Vec<bool> = Vec::with_capacity(ctx.num_rows);
-        let mut current_offset: i32 = 0;
+        let mut current_offset: i32 = 0;  // TODO: consider switching to i64
 
         // Pre-compute the mini-batch schema once - it's the same for all rows.
         let mini_batch_schema =

datafusion/bio-function-vcftools/tests/integration_test.rs

@@ -971,7 +971,7 @@ async fn test_mismatched_array_lengths() {
     let df_optimized2 = ctx_optimized.sql(sql).await.unwrap();
     let optimized_results = df_optimized2.collect().await.unwrap();
 
-    // Both baseline and optimized should filter out the empty-array row and have identical output
+    // Both baseline and optimized should handle mismatched lengths gracefully and produce identical output
     let baseline_str = pretty_format_batches(&baseline_results)
         .unwrap()
         .to_string();