fpetkovski commented on code in PR #22809: URL: https://github.com/apache/datafusion/pull/22809#discussion_r3373496507
########## datafusion-examples/examples/data_io/partitioned_file_schema.rs: ########## @@ -0,0 +1,140 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +//! See `main.rs` for how to run it. + +use arrow::array::{Int32Array, RecordBatch}; +use arrow_schema::{DataType, Field, Schema, SchemaRef}; +use datafusion::common::Result; +use datafusion::datasource::listing::PartitionedFile; +use datafusion::datasource::object_store::ObjectStoreUrl; +use datafusion::datasource::physical_plan::{FileScanConfigBuilder, ParquetSource}; +use datafusion::datasource::source::DataSourceExec; +use datafusion::execution::TaskContext; +use datafusion::parquet::arrow::ArrowWriter; +use datafusion::parquet::file::reader::Length; +use datafusion::physical_plan::ExecutionPlan; +use std::fs::File; +use std::path::Path; +use std::sync::Arc; +use tempfile::TempDir; +use tonic::codegen::tokio_stream::StreamExt; + +/// Demonstrates how to attach a per-file Arrow schema to a [`PartitionedFile`] +/// via [`PartitionedFile::with_arrow_schema`]. +/// +/// By default DataFusion infers a file's physical schema by reading its +/// metadata (e.g. the Parquet footer) when the scan begins. When the schema is +/// already known, it can be supplied up front so this inference step is +/// skipped, saving an I/O round trip and metadata parse per file. +/// +/// The example writes a small Parquet file with a single `Int32` column `a` and +/// reads it back three ways: +/// - without a schema, letting DataFusion infer it at query time; +/// - with the correct schema, skipping inference; +/// - with a deliberately mismatched schema (`a` typed as `Int64`), which +/// surfaces as an error since the provided schema does not match the data +/// actually stored in the file. +/// +/// Note that the schema passed to [`PartitionedFile::with_arrow_schema`] must +/// describe only the columns physically stored in the file and must not include +/// partition columns. +pub async fn read_partitioned_file() -> Result<()> { + let tmpdir = TempDir::new()?; + let file_path = tmpdir.path().join("partitioned-file"); + let file = File::create(file_path.as_path())?; + + let file_schema = + Arc::new(Schema::new(vec![Field::new("a", DataType::Int32, false)])); + let batch = RecordBatch::try_new( + file_schema.clone(), + vec![Arc::new(Int32Array::from(vec![1, 2, 3, 4, 5]))], + )?; + let mut writer = ArrowWriter::try_new(&file, file_schema.clone(), None)?; + writer.write(&batch)?; + writer.finish()?; + + let table_schema = Arc::new(Schema::new(vec![ + Field::new("a", DataType::Int32, true), + Field::new("b", DataType::Float64, true), Review Comment: I added a comment for this field, let me know if it makes sense. ########## datafusion-examples/examples/data_io/main.rs: ########## Review Comment: Thanks, updated. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
