This is an automated email from the ASF dual-hosted git repository.

comphead pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/datafusion.git


The following commit(s) were added to refs/heads/main by this push:
     new 8f76ac553a Docs: clarify when the reader will read from object store 
when using cached metadata (#10909)
8f76ac553a is described below

commit 8f76ac553a4a05705cb92eb4fe46158b4bfa354b
Author: Andrew Lamb <[email protected]>
AuthorDate: Fri Jun 14 11:54:31 2024 -0400

    Docs: clarify when the reader will read from object store when using cached 
metadata (#10909)
---
 .../core/src/datasource/physical_plan/parquet/reader.rs       | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/datafusion/core/src/datasource/physical_plan/parquet/reader.rs 
b/datafusion/core/src/datasource/physical_plan/parquet/reader.rs
index 265fb9d570..8a4ba136fc 100644
--- a/datafusion/core/src/datasource/physical_plan/parquet/reader.rs
+++ b/datafusion/core/src/datasource/physical_plan/parquet/reader.rs
@@ -16,7 +16,7 @@
 // under the License.
 
 //! [`ParquetFileReaderFactory`] and [`DefaultParquetFileReaderFactory`] for
-//! creating parquet file readers
+//! low level control of parquet file readers
 
 use crate::datasource::physical_plan::{FileMeta, ParquetFileMetrics};
 use bytes::Bytes;
@@ -33,12 +33,19 @@ use std::sync::Arc;
 ///
 /// The combined implementations of [`ParquetFileReaderFactory`] and
 /// [`AsyncFileReader`] can be used to provide custom data access operations
-/// such as pre-cached data, I/O coalescing, etc.
+/// such as pre-cached metadata, I/O coalescing, etc.
 ///
 /// See [`DefaultParquetFileReaderFactory`] for a simple implementation.
 pub trait ParquetFileReaderFactory: Debug + Send + Sync + 'static {
     /// Provides an `AsyncFileReader` for reading data from a parquet file 
specified
     ///
+    /// # Notes
+    ///
+    /// If the resulting [`AsyncFileReader`]  returns `ParquetMetaData` without
+    /// page index information, the reader will load it on demand. Thus it is 
important
+    /// to ensure that the returned `ParquetMetaData` has the necessary 
information
+    /// if you wish to avoid a subsequent I/O
+    ///
     /// # Arguments
     /// * partition_index - Index of the partition (for reporting metrics)
     /// * file_meta - The file to be read


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to