ARROW-4219: [Rust] [Parquet] Initial support for arrow reader.

rust/arrow/src/array/array.rs

@@ -1056,6 +1056,11 @@ impl StructArray {
         self.boxed_fields.iter().collect()
     }
 
+    /// Returns child array refs of the struct array
+    pub fn columns_ref(&self) -> Vec<ArrayRef> {
+        self.boxed_fields.clone()
+    }
+
     /// Return field names in this struct array
     pub fn column_names(&self) -> Vec<&str> {
         match self.data.data_type() {

rust/arrow/src/datatypes.rs

@@ -34,6 +34,7 @@ use serde_derive::{Deserialize, Serialize};
 use serde_json::{json, Number, Value, Value::Number as VNumber};
 
 use crate::error::{ArrowError, Result};
+use std::sync::Arc;
 
 /// The possible relative types that are supported.
 ///
@@ -869,6 +870,8 @@ impl fmt::Display for Schema {
     }
 }
 
+pub type SchemaRef = Arc<Schema>;
+
 #[cfg(test)]
 mod tests {
     use super::*;

rust/arrow/src/error.rs

@@ -31,6 +31,7 @@ pub enum ArrowError {
     JsonError(String),
     IoError(String),
     InvalidArgumentError(String),
+    ParquetError(String),
 }
 
 impl From<::std::io::Error> for ArrowError {

rust/arrow/src/record_batch.rs

@@ -113,9 +113,32 @@ impl From<&StructArray> for RecordBatch {
     }
 }
 
+impl Into<StructArray> for RecordBatch {
+    fn into(self) -> StructArray {
+        self.schema
+            .fields
+            .iter()
+            .zip(self.columns.iter())
+            .map(|t| (t.0.clone(), t.1.clone()))
+            .collect::<Vec<(Field, ArrayRef)>>()
+            .into()
+    }
+}
+
 unsafe impl Send for RecordBatch {}
 unsafe impl Sync for RecordBatch {}
 
+/// Definition of record batch reader.
+pub trait RecordBatchReader {
+    /// Returns schemas of this record batch reader.
+    /// Implementation of this trait should guarantee that all record batches returned
+    /// by this reader should have same schema as returned from this method.
+    fn schema(&mut self) -> SchemaRef;
+
+    /// Returns next record batch.
+    fn next_batch(&mut self) -> Result<Option<RecordBatch>>;
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;

rust/parquet/Cargo.toml

@@ -41,6 +41,7 @@ zstd = "0.4"
 chrono = "0.4"
 num-bigint = "0.2"
 arrow = { path = "../arrow", version = "1.0.0-SNAPSHOT" }
+serde_json = { version = "1.0.13", features = ["preserve_order"] }
 
 [dev-dependencies]
 lazy_static = "1"

rust/parquet/src/arrow/array_reader.rs

@@ -52,9 +52,12 @@ use crate::schema::types::{
     ColumnDescPtr, ColumnDescriptor, ColumnPath, SchemaDescPtr, Type, TypePtr,
 };
 use crate::schema::visitor::TypeVisitor;
+use std::any::Any;
 
 /// Array reader reads parquet data into arrow array.
 pub trait ArrayReader {
+    fn as_any(&self) -> &dyn Any;
+
     /// Returns the arrow type of this array reader.
     fn get_data_type(&self) -> &ArrowType;
 
@@ -115,6 +118,10 @@ impl<T: DataType> PrimitiveArrayReader<T> {
 
 /// Implementation of primitive array reader.
 impl<T: DataType> ArrayReader for PrimitiveArrayReader<T> {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
     /// Returns data type of primitive array.
     fn get_data_type(&self) -> &ArrowType {
         &self.data_type
@@ -232,7 +239,7 @@ impl<T: DataType> ArrayReader for PrimitiveArrayReader<T> {
 }
 
 /// Implementation of struct array reader.
-struct StructArrayReader {
+pub struct StructArrayReader {
     children: Vec<Box<dyn ArrayReader>>,
     data_type: ArrowType,
     struct_def_level: i16,
@@ -261,6 +268,10 @@ impl StructArrayReader {
 }
 
 impl ArrayReader for StructArrayReader {
+    fn as_any(&self) -> &dyn Any {
+        self
+    }
+
     /// Returns data type.
     /// This must be a struct.
     fn get_data_type(&self) -> &ArrowType {
@@ -705,6 +716,7 @@ mod tests {
     use arrow::array::{Array, ArrayRef, PrimitiveArray, StructArray};
     use arrow::datatypes::{DataType as ArrowType, Field, Int32Type as ArrowInt32};
     use rand::distributions::range::SampleRange;
+    use std::any::Any;
     use std::collections::VecDeque;
     use std::rc::Rc;
     use std::sync::Arc;
@@ -953,6 +965,10 @@ mod tests {
     }
 
     impl ArrayReader for InMemoryArrayReader {
+        fn as_any(&self) -> &Any {
+            self
+        }
+
         fn get_data_type(&self) -> &ArrowType {
             &self.data_type
         }