Upgrade to arrow/parquet 55, and object_store to 0.12.0 and pyo3 to 0.24.0 (apache#15466)

* Temp pin to datafusion main

* Update cargo lock

* update pyo3

* vendor random generation

* Update error message

* Update for extraction

* Update pin

* Upgrade object_store

* fix feature

* Update file size handling

* bash for object store API changes

* few more

* Update APIs more

* update expected message

* update error messages

* Update to apache

* Update API for nicer parquet u64s

* Fix wasm build

* Remove pin

* Fix signature

Author: alamb

Cargo.lock

@@ -87,19 +87,19 @@ ahash = { version = "0.8", default-features = false, features = [
     "runtime-rng",
 ] }
 apache-avro = { version = "0.17", default-features = false }
-arrow = { version = "54.3.1", features = [
+arrow = { version = "55.0.0", features = [
     "prettyprint",
     "chrono-tz",
 ] }
-arrow-buffer = { version = "54.3.0", default-features = false }
-arrow-flight = { version = "54.3.1", features = [
+arrow-buffer = { version = "55.0.0", default-features = false }
+arrow-flight = { version = "55.0.0", features = [
     "flight-sql-experimental",
 ] }
-arrow-ipc = { version = "54.3.0", default-features = false, features = [
+arrow-ipc = { version = "55.0.0", default-features = false, features = [
     "lz4",
 ] }
-arrow-ord = { version = "54.3.0", default-features = false }
-arrow-schema = { version = "54.3.0", default-features = false }
+arrow-ord = { version = "55.0.0", default-features = false }
+arrow-schema = { version = "55.0.0", default-features = false }
 async-trait = "0.1.88"
 bigdecimal = "0.4.8"
 bytes = "1.10"
@@ -147,9 +147,9 @@ hashbrown = { version = "0.14.5", features = ["raw"] }
 indexmap = "2.8.0"
 itertools = "0.14"
 log = "^0.4"
-object_store = { version = "0.11.0", default-features = false }
+object_store = { version = "0.12.0", default-features = false }
 parking_lot = "0.12"
-parquet = { version = "54.3.1", default-features = false, features = [
+parquet = { version = "55.0.0", default-features = false, features = [
     "arrow",
     "async",
     "object_store",

Cargo.toml

@@ -571,7 +571,9 @@ impl ParquetFileReaderFactory for CachedParquetFileReaderFactory {
             .to_string();
 
         let object_store = Arc::clone(&self.object_store);
-        let mut inner = ParquetObjectReader::new(object_store, file_meta.object_meta);
+        let mut inner =
+            ParquetObjectReader::new(object_store, file_meta.object_meta.location)
+                .with_file_size(file_meta.object_meta.size);
 
         if let Some(hint) = metadata_size_hint {
             inner = inner.with_footer_size_hint(hint)
@@ -599,15 +601,15 @@ struct ParquetReaderWithCache {
 impl AsyncFileReader for ParquetReaderWithCache {
     fn get_bytes(
         &mut self,
-        range: Range<usize>,
+        range: Range<u64>,
     ) -> BoxFuture<'_, datafusion::parquet::errors::Result<Bytes>> {
         println!("get_bytes: {} Reading range {:?}", self.filename, range);
         self.inner.get_bytes(range)
     }
 
     fn get_byte_ranges(
         &mut self,
-        ranges: Vec<Range<usize>>,
+        ranges: Vec<Range<u64>>,
     ) -> BoxFuture<'_, datafusion::parquet::errors::Result<Vec<Bytes>>> {
         println!(
             "get_byte_ranges: {} Reading ranges {:?}",
@@ -618,6 +620,7 @@ impl AsyncFileReader for ParquetReaderWithCache {
 
     fn get_metadata(
         &mut self,
+        _options: Option<&ArrowReaderOptions>,
     ) -> BoxFuture<'_, datafusion::parquet::errors::Result<Arc<ParquetMetaData>>> {
         println!("get_metadata: {} returning cached metadata", self.filename);
 

datafusion-examples/examples/advanced_parquet_index.rs

@@ -63,7 +63,7 @@ log = { workspace = true }
 object_store = { workspace = true, optional = true }
 parquet = { workspace = true, optional = true, default-features = true }
 paste = "1.0.15"
-pyo3 = { version = "0.23.5", optional = true }
+pyo3 = { version = "0.24.0", optional = true }
 recursive = { workspace = true, optional = true }
 sqlparser = { workspace = true }
 tokio = { workspace = true }

datafusion/common/Cargo.toml

@@ -144,6 +144,7 @@ impl FileFormat for ArrowFormat {
         for object in objects {
             let r = store.as_ref().get(&object.location).await?;
             let schema = match r.payload {
+                #[cfg(not(target_arch = "wasm32"))]
                 GetResultPayload::File(mut file, _) => {
                     let reader = FileReader::try_new(&mut file, None)?;
                     reader.schema()
@@ -442,7 +443,7 @@ mod tests {
         let object_meta = ObjectMeta {
             location,
             last_modified: DateTime::default(),
-            size: usize::MAX,
+            size: u64::MAX,
             e_tag: None,
             version: None,
         };
@@ -485,7 +486,7 @@ mod tests {
         let object_meta = ObjectMeta {
             location,
             last_modified: DateTime::default(),
-            size: usize::MAX,
+            size: u64::MAX,
             e_tag: None,
             version: None,
         };

datafusion/core/src/datasource/file_format/arrow.rs

@@ -72,7 +72,7 @@ mod tests {
     #[derive(Debug)]
     struct VariableStream {
         bytes_to_repeat: Bytes,
-        max_iterations: usize,
+        max_iterations: u64,
         iterations_detected: Arc<Mutex<usize>>,
     }
 
@@ -103,14 +103,15 @@ mod tests {
 
         async fn get(&self, location: &Path) -> object_store::Result<GetResult> {
             let bytes = self.bytes_to_repeat.clone();
-            let range = 0..bytes.len() * self.max_iterations;
+            let len = bytes.len() as u64;
+            let range = 0..len * self.max_iterations;
             let arc = self.iterations_detected.clone();
             let stream = futures::stream::repeat_with(move || {
                 let arc_inner = arc.clone();
                 *arc_inner.lock().unwrap() += 1;
                 Ok(bytes.clone())
             })
-            .take(self.max_iterations)
+            .take(self.max_iterations as usize)
             .boxed();
 
             Ok(GetResult {
@@ -138,7 +139,7 @@ mod tests {
         async fn get_ranges(
             &self,
             _location: &Path,
-            _ranges: &[Range<usize>],
+            _ranges: &[Range<u64>],
         ) -> object_store::Result<Vec<Bytes>> {
             unimplemented!()
         }
@@ -154,7 +155,7 @@ mod tests {
         fn list(
             &self,
             _prefix: Option<&Path>,
-        ) -> BoxStream<'_, object_store::Result<ObjectMeta>> {
+        ) -> BoxStream<'static, object_store::Result<ObjectMeta>> {
             unimplemented!()
         }
 
@@ -179,7 +180,7 @@ mod tests {
     }
 
     impl VariableStream {
-        pub fn new(bytes_to_repeat: Bytes, max_iterations: usize) -> Self {
+        pub fn new(bytes_to_repeat: Bytes, max_iterations: u64) -> Self {
             Self {
                 bytes_to_repeat,
                 max_iterations,
@@ -371,7 +372,7 @@ mod tests {
         let object_meta = ObjectMeta {
             location: Path::parse("/")?,
             last_modified: DateTime::default(),
-            size: usize::MAX,
+            size: u64::MAX,
             e_tag: None,
             version: None,
         };
@@ -429,7 +430,7 @@ mod tests {
         let object_meta = ObjectMeta {
             location: Path::parse("/")?,
             last_modified: DateTime::default(),
-            size: usize::MAX,
+            size: u64::MAX,
             e_tag: None,
             version: None,
         };

datafusion/core/src/datasource/file_format/csv.rs

@@ -127,7 +127,7 @@ mod tests {
             .write_parquet(out_dir_url, DataFrameWriteOptions::new(), None)
             .await
             .expect_err("should fail because input file does not match inferred schema");
-        assert_eq!(e.strip_backtrace(), "Arrow error: Parser error: Error while parsing value d for column 0 at line 4");
+        assert_eq!(e.strip_backtrace(), "Arrow error: Parser error: Error while parsing value 'd' as type 'Int64' for column 0 at line 4. Row data: '[d,4]'");
         Ok(())
     }
 }

datafusion/core/src/datasource/file_format/mod.rs

@@ -331,7 +331,7 @@ mod tests {
         fn list(
             &self,
             _prefix: Option<&Path>,
-        ) -> BoxStream<'_, object_store::Result<ObjectMeta>> {
+        ) -> BoxStream<'static, object_store::Result<ObjectMeta>> {
             Box::pin(futures::stream::once(async {
                 Err(object_store::Error::NotImplemented)
             }))
@@ -408,7 +408,7 @@ mod tests {
         )));
 
         // Use the file size as the hint so we can get the full metadata from the first fetch
-        let size_hint = meta[0].size;
+        let size_hint = meta[0].size as usize;
 
         fetch_parquet_metadata(store.upcast().as_ref(), &meta[0], Some(size_hint))
             .await
@@ -443,7 +443,7 @@ mod tests {
         )));
 
         // Use the a size hint larger than the file size to make sure we don't panic
-        let size_hint = meta[0].size + 100;
+        let size_hint = (meta[0].size + 100) as usize;
 
         fetch_parquet_metadata(store.upcast().as_ref(), &meta[0], Some(size_hint))
             .await

datafusion/core/src/datasource/file_format/parquet.rs

@@ -106,7 +106,7 @@ mod tests {
         let meta = ObjectMeta {
             location,
             last_modified: metadata.modified().map(chrono::DateTime::from).unwrap(),
-            size: metadata.len() as usize,
+            size: metadata.len(),
             e_tag: None,
             version: None,
         };

datafusion/core/src/datasource/mod.rs

@@ -273,6 +273,7 @@ impl FileOpener for ArrowOpener {
                 None => {
                     let r = object_store.get(file_meta.location()).await?;
                     match r.payload {
+                        #[cfg(not(target_arch = "wasm32"))]
                         GetResultPayload::File(file, _) => {
                             let arrow_reader = arrow::ipc::reader::FileReader::try_new(
                                 file, projection,
@@ -305,7 +306,7 @@ impl FileOpener for ArrowOpener {
                     )?;
                     // read footer according to footer_len
                     let get_option = GetOptions {
-                        range: Some(GetRange::Suffix(10 + footer_len)),
+                        range: Some(GetRange::Suffix(10 + (footer_len as u64))),
                         ..Default::default()
                     };
                     let get_result = object_store
@@ -332,9 +333,9 @@ impl FileOpener for ArrowOpener {
                         .iter()
                         .flatten()
                         .map(|block| {
-                            let block_len = block.bodyLength() as usize
-                                + block.metaDataLength() as usize;
-                            let block_offset = block.offset() as usize;
+                            let block_len =
+                                block.bodyLength() as u64 + block.metaDataLength() as u64;
+                            let block_offset = block.offset() as u64;
                             block_offset..block_offset + block_len
                         })
                         .collect_vec();
@@ -354,19 +355,19 @@ impl FileOpener for ArrowOpener {
                         .iter()
                         .flatten()
                         .filter(|block| {
-                            let block_offset = block.offset() as usize;
-                            block_offset >= range.start as usize
-                                && block_offset < range.end as usize
+                            let block_offset = block.offset() as u64;
+                            block_offset >= range.start as u64
+                                && block_offset < range.end as u64
                         })
                         .copied()
                         .collect_vec();
 
                     let recordbatch_ranges = recordbatches
                         .iter()
                         .map(|block| {
-                            let block_len = block.bodyLength() as usize
-                                + block.metaDataLength() as usize;
-                            let block_offset = block.offset() as usize;
+                            let block_len =
+                                block.bodyLength() as u64 + block.metaDataLength() as u64;
+                            let block_offset = block.offset() as u64;
                             block_offset..block_offset + block_len
                         })
                         .collect_vec();