Fix up doc tests

Author: friendlymatthew

datafusion/datasource-parquet/src/source.rs

@@ -98,18 +98,20 @@ use object_store::ObjectStore;
 /// # use datafusion_physical_expr::expressions::lit;
 /// # use datafusion_datasource::source::DataSourceExec;
 /// # use datafusion_common::config::TableParquetOptions;
+/// use datafusion_datasource::file::FileSource;
 ///
 /// # let file_schema = Arc::new(Schema::empty());
 /// # let object_store_url = ObjectStoreUrl::local_filesystem();
 /// # let predicate = lit(true);
-/// let source = Arc::new(
-///     ParquetSource::new(TableParquetOptions::default(), config)
-///     .with_predicate(predicate)
-/// );
 /// // Create a DataSourceExec for reading `file1.parquet` with a file size of 100MB
-/// let config = FileScanConfigBuilder::new(object_store_url, file_schema, source)
+/// let config = FileScanConfigBuilder::new(object_store_url, file_schema)
 ///    .with_file(PartitionedFile::new("file1.parquet", 100*1024*1024)).build();
-/// let exec =DataSourceExec::from_data_source(source);
+///
+/// let source = ParquetSource::new(TableParquetOptions::default(), config)
+///     .with_predicate(predicate)
+///     .as_data_source();
+///
+/// let exec = Arc::new(DataSourceExec::new(source));
 /// ```
 ///
 /// # Features
@@ -176,7 +178,7 @@ use object_store::ObjectStore;
 /// # use datafusion_datasource::file::FileSource;
 /// # use datafusion_datasource::file_scan_config::{FileScanConfig, FileScanConfigBuilder};
 /// # use datafusion_datasource::PartitionedFile;
-/// # use datafusion_datasource::source::DataSourceExec;///
+/// # use datafusion_datasource::source::DataSourceExec;
 /// # use datafusion_datasource_parquet::source::ParquetSource;
 ///
 /// fn parquet_exec() -> DataSourceExec { unimplemented!() }
@@ -194,7 +196,9 @@ use object_store::ObjectStore;
 ///        .with_file_groups(vec![file_group.clone()])
 ///       .build();
 ///
-///     Arc::new(DataSourceExec::new(parquet_source.with_config(new_config)))
+///     let source = parquet_source.with_config(new_config).as_data_source();
+///
+///     Arc::new(DataSourceExec::new(source))
 ///   })
 ///   .collect::<Vec<_>>();
 /// ```
@@ -215,6 +219,7 @@ use object_store::ObjectStore;
 /// ```
 /// # use std::sync::Arc;
 /// # use arrow::datatypes::{Schema, SchemaRef};
+/// # use datafusion_common::config::TableParquetOptions;
 /// # use datafusion_datasource::PartitionedFile;
 /// # use datafusion_datasource_parquet::ParquetAccessPlan;
 /// # use datafusion_datasource::file_scan_config::{FileScanConfig, FileScanConfigBuilder};
@@ -233,11 +238,14 @@ use object_store::ObjectStore;
 /// let partitioned_file = PartitionedFile::new("my_file.parquet", 1234)
 ///   .with_extensions(Arc::new(access_plan));
 /// // create a FileScanConfig to scan this file
-/// let config = FileScanConfigBuilder::new(ObjectStoreUrl::local_filesystem(), schema(), Arc::new(ParquetSource::new(TableParquetOptions::default(), config)))
+/// let config = FileScanConfigBuilder::new(ObjectStoreUrl::local_filesystem(), schema())
 ///     .with_file(partitioned_file).build();
+///
+/// let source = ParquetSource::new(TableParquetOptions::default(), config);
+///
 /// // this parquet DataSourceExec will not even try to read row groups 2 and 4. Additional
 /// // pruning based on predicates may also happen
-/// let exec =DataSourceExec::from_data_source(source);
+/// let exec = DataSourceExec::from_data_source(source);
 /// ```
 ///
 /// For a complete example, see the [`advanced_parquet_index` example]).

datafusion/datasource/src/file_scan_config.rs

@@ -60,6 +60,8 @@ use log::{debug, warn};
 /// Owned by file sources that implement [`DataSource`]. Use [`DataSourceExec::from_data_source`]
 /// to create a [`DataSourceExec`] from a data source that owns this config.
 ///
+/// [`DataSourceExec::from_data_source`]: crate::source::DataSourceExec::from_data_source
+/// [`DataSourceExec`]: crate::source::DataSourceExec
 /// [`DataSource`]: crate::source::DataSource
 ///
 /// # Example
@@ -237,7 +239,7 @@ pub struct FileScanConfigBuilder {
     /// This schema is used to read the files, but is **not** necessarily the
     /// schema of the physical files. Rather this is the schema that the
     /// physical file schema will be mapped onto, and the schema that the
-    /// [`DataSourceExec`] will return.
+    /// [`crate::source::DataSourceExec`] will return.
     ///
     /// This is usually the same as the table schema as specified by the `TableProvider` minus any partition columns.
     ///

datafusion/datasource/src/source.rs

@@ -468,7 +468,7 @@ impl DataSourceExec {
     ///
     /// Returns `None` if
     /// 1. the datasource is not scanning files (`FileScanConfig`)
-    /// 2. The [`FileScanConfig::file_source`] is not of type `T`
+    /// 2. the file source is not of type `T`
     pub fn downcast_to_file_source<T: FileSource + 'static>(&self) -> Option<&T> {
         self.data_source().as_any().downcast_ref::<T>()
     }