Only generate data if looking for local data

Signed-off-by: Adam Gutglick <[email protected]>

Author: AdamGS

.github/workflows/sql-benchmarks.yml

@@ -246,7 +246,7 @@ jobs:
           OTEL_EXPORTER_OTLP_ENDPOINT: "${{ (inputs.mode != 'pr' || github.event.pull_request.head.repo.fork == false) && secrets.OTEL_EXPORTER_OTLP_ENDPOINT || '' }}"
           OTEL_EXPORTER_OTLP_HEADERS: "${{ (inputs.mode != 'pr' || github.event.pull_request.head.repo.fork == false) && secrets.OTEL_EXPORTER_OTLP_HEADERS || '' }}"
           OTEL_RESOURCE_ATTRIBUTES: "bench-name=${{ matrix.id }}"
-          RUST_LOG: "RUST_LOG=debug,vortex_io=off,hyper=off,datafusion=info,reqwest=off,sqlparser=off"
+          RUST_LOG: "RUST_LOG=debug,vortex_io=off,hyper=off,datafusion=info,sqlparser=off"
         run: |
           # Extract formats for each engine
           df_formats=$(echo "${{ matrix.targets }}" | tr ',' '\n' | grep '^datafusion:' | sed 's/datafusion://' | tr '\n' ',' | sed 's/,$//')

benchmarks/ddb-bench/src/main.rs

@@ -83,11 +83,10 @@ async fn main() -> anyhow::Result<()> {
         args.exclude_queries.as_ref(),
     );
 
-    // Generate base Parquet data first - all formats derive from this
-    benchmark.generate_base_data().await?;
-
     // Generate Vortex files from Parquet for any Vortex formats requested
     if benchmark.data_url().scheme() == "file" {
+        benchmark.generate_base_data().await?;
+
         let base_path = benchmark
             .data_url()
             .to_file_path()

benchmarks/df-bench/src/main.rs

@@ -109,10 +109,10 @@ async fn main() -> anyhow::Result<()> {
         args.exclude_queries.as_ref(),
     );
 
-    benchmark.generate_base_data().await?;
-
     // Generate Vortex files from Parquet for any Vortex formats requested
     if benchmark.data_url().scheme() == "file" {
+        benchmark.generate_base_data().await?;
+
         let base_path = benchmark
             .data_url()
             .to_file_path()