Add Row size methods (apache#3160)

arrow/src/lib.rs

@@ -297,6 +297,8 @@
 #![deny(clippy::redundant_clone)]
 #![warn(missing_debug_implementations)]
 #![allow(rustdoc::invalid_html_tags)]
+extern crate core;
+
 pub use arrow_array::{downcast_dictionary_array, downcast_primitive_array};
 
 pub use arrow_buffer::{alloc, buffer};

arrow/src/row/interner.rs

@@ -157,6 +157,15 @@ impl OrderPreservingInterner {
     pub fn value(&self, key: Interned) -> &[u8] {
         self.values.index(key)
     }
+
+    /// Returns the size of this instance in bytes including self
+    pub fn size(&self) -> usize {
+        std::mem::size_of::<Self>()
+            + self.values.buffer_size()
+            + self.values.buffer_size()
+            + self.bucket.size()
+            + self.lookup.capacity() * std::mem::size_of::<Interned>()
+    }
 }
 
 /// A buffer of `[u8]` indexed by `[Interned]`
@@ -192,6 +201,11 @@ impl InternBuffer {
         self.offsets.push(self.values.len());
         key
     }
+
+    /// Returns the byte size of the associated buffers
+    fn buffer_size(&self) -> usize {
+        self.values.capacity() + self.offsets.capacity() * std::mem::size_of::<usize>()
+    }
 }
 
 impl Index<Interned> for InternBuffer {
@@ -324,6 +338,13 @@ impl Bucket {
             }
         }
     }
+
+    /// Returns the size of this instance in bytes
+    fn size(&self) -> usize {
+        std::mem::size_of::<Self>()
+            + self.slots.capacity() * std::mem::size_of::<Slot>()
+            + self.next.as_ref().map(|x| x.size()).unwrap_or_default()
+    }
 }
 
 #[cfg(test)]

arrow/src/row/mod.rs

@@ -128,6 +128,7 @@
 //! [compared]: PartialOrd
 //! [compare]: PartialOrd
 
+use core::slice::SlicePattern;
 use std::cmp::Ordering;
 use std::hash::{Hash, Hasher};
 use std::sync::Arc;
@@ -358,6 +359,14 @@ impl SortField {
     pub fn new_with_options(data_type: DataType, options: SortOptions) -> Self {
         Self { options, data_type }
     }
+
+    /// Return size of this instance in bytes.
+    ///
+    /// Includes the size of `Self`.
+    pub fn size(&self) -> usize {
+        self.data_type.size() + std::mem::size_of::<Self>()
+            - std::mem::size_of::<DataType>()
+    }
 }
 
 impl RowConverter {
@@ -480,6 +489,22 @@ impl RowConverter {
             })
             .collect()
     }
+
+    /// Returns the size of this instance in bytes
+    ///
+    /// Includes the size of `Self`.
+    pub fn size(&self) -> usize {
+        std::mem::size_of::<Self>()
+            + std::mem::size_of_val(&self.interners)
+            + self.fields.iter().map(|x| x.size()).sum::<usize>()
+            + self.interners.capacity()
+                * std::mem::size_of::<Option<Box<OrderPreservingInterner>>>()
+            + self
+                .interners
+                .iter()
+                .filter_map(|x| x.as_ref().map(|x| x.size()))
+                .sum::<usize>()
+    }
 }
 
 /// A row-oriented representation of arrow data, that is normalized for comparison.
@@ -512,6 +537,15 @@ impl Rows {
     pub fn iter(&self) -> RowsIter<'_> {
         self.into_iter()
     }
+
+    /// Returns the size of this instance in bytes
+    ///
+    /// Includes the size of `Self`.
+    pub fn size(&self) -> usize {
+        std::mem::size_of::<Self>()
+            + std::mem::size_of_val(self.buffer.as_slice())
+            + std::mem::size_of_val(self.offsets.as_slice())
+    }
 }
 
 impl<'a> IntoIterator for &'a Rows {