change ColumnChunkMetaData::encodings to return an iterator as

suggested in review.

Author: etseidl

parquet/src/basic.rs

@@ -769,11 +769,10 @@ where
     mask
 }
 
-pub(super) fn mask_to_encodings_vec(mask: i32) -> Vec<Encoding> {
+pub(super) fn mask_to_encodings_iter(mask: i32) -> impl Iterator<Item = Encoding> {
     (0..=MAX_ENCODING)
-        .filter(|i| mask & (1 << i) != 0)
+        .filter(move |i| mask & (1 << i) != 0)
         .map(i32_to_encoding)
-        .collect()
 }
 
 // ----------------------------------------------------------------------
@@ -2465,7 +2464,10 @@ mod tests {
     fn encodings_roundtrip(encodings: &mut [Encoding]) {
         encodings.sort();
         let mask = encodings_to_mask(encodings.iter());
-        assert_eq!(mask_to_encodings_vec(mask), encodings.to_vec());
+        assert_eq!(
+            mask_to_encodings_iter(mask).collect::<Vec<_>>(),
+            encodings.to_vec()
+        );
     }
 
     #[test]

parquet/src/column/writer/mod.rs

@@ -1735,7 +1735,10 @@ mod tests {
         assert_eq!(r.rows_written, 4);
 
         let metadata = r.metadata;
-        assert_eq!(metadata.encodings(), vec![Encoding::PLAIN, Encoding::RLE]);
+        assert_eq!(
+            metadata.encodings().collect::<Vec<_>>(),
+            vec![Encoding::PLAIN, Encoding::RLE]
+        );
         assert_eq!(metadata.num_values(), 4); // just values
         assert_eq!(metadata.dictionary_page_offset(), None);
     }
@@ -2096,7 +2099,7 @@ mod tests {
 
         let metadata = r.metadata;
         assert_eq!(
-            metadata.encodings(),
+            metadata.encodings().collect::<Vec<_>>(),
             vec![Encoding::PLAIN, Encoding::RLE, Encoding::RLE_DICTIONARY]
         );
         assert_eq!(metadata.num_values(), 4);
@@ -2222,7 +2225,7 @@ mod tests {
 
         let metadata = r.metadata;
         assert_eq!(
-            metadata.encodings(),
+            metadata.encodings().collect::<Vec<_>>(),
             vec![Encoding::PLAIN, Encoding::RLE, Encoding::RLE_DICTIONARY]
         );
         assert_eq!(metadata.num_values(), 4);
@@ -4101,7 +4104,7 @@ mod tests {
             .build();
         let meta = column_write_and_get_metadata::<T>(props, data);
         assert_eq!(meta.dictionary_page_offset(), dictionary_page_offset);
-        assert_eq!(meta.encodings(), encodings);
+        assert_eq!(meta.encodings().collect::<Vec<_>>(), encodings);
         assert_eq!(meta.page_encoding_stats().unwrap(), page_encoding_stats);
     }
 

parquet/src/file/metadata/mod.rs

@@ -97,7 +97,7 @@ pub(crate) mod reader;
 pub(crate) mod thrift_gen;
 mod writer;
 
-use crate::basic::{PageType, encodings_to_mask, mask_to_encodings_vec};
+use crate::basic::{PageType, encodings_to_mask, mask_to_encodings_iter};
 #[cfg(feature = "encryption")]
 use crate::encryption::decrypt::FileDecryptor;
 #[cfg(feature = "encryption")]
@@ -970,8 +970,8 @@ impl ColumnChunkMetaData {
     }
 
     /// All encodings used for this column.
-    pub fn encodings(&self) -> Vec<Encoding> {
-        mask_to_encodings_vec(self.encodings)
+    pub fn encodings(&self) -> impl Iterator<Item = Encoding> {
+        mask_to_encodings_iter(self.encodings)
     }
 
     /// All encodings used for this column, returned as a bitmask.

parquet/src/file/metadata/thrift_gen.rs

@@ -1262,7 +1262,10 @@ pub(crate) fn serialize_column_meta_data<W: Write>(
     use crate::file::statistics::page_stats_to_thrift;
 
     column_chunk.column_type().write_thrift_field(w, 1, 0)?;
-    column_chunk.encodings().write_thrift_field(w, 2, 1)?;
+    column_chunk
+        .encodings()
+        .collect::<Vec<_>>()
+        .write_thrift_field(w, 2, 1)?;
     let path = column_chunk.column_descr.path().parts();
     let path: Vec<&str> = path.iter().map(|v| v.as_str()).collect();
     path.write_thrift_field(w, 3, 2)?;

parquet/src/file/writer.rs

@@ -2390,6 +2390,7 @@ mod tests {
                     .row_group(0)
                     .column(x)
                     .encodings()
+                    .collect::<Vec<_>>()
                     .contains(&Encoding::BYTE_STREAM_SPLIT)
             );
         };

parquet/src/schema/printer.rs

@@ -171,11 +171,7 @@ fn print_row_group_metadata(out: &mut dyn io::Write, rg_metadata: &RowGroupMetaD
 fn print_column_chunk_metadata(out: &mut dyn io::Write, cc_metadata: &ColumnChunkMetaData) {
     writeln!(out, "column type: {}", cc_metadata.column_type());
     writeln!(out, "column path: {}", cc_metadata.column_path());
-    let encoding_strs: Vec<_> = cc_metadata
-        .encodings()
-        .iter()
-        .map(|e| format!("{e}"))
-        .collect();
+    let encoding_strs: Vec<_> = cc_metadata.encodings().map(|e| format!("{e}")).collect();
     writeln!(out, "encodings: {}", encoding_strs.join(" "));
     let file_path_str = cc_metadata.file_path().unwrap_or("N/A");
     writeln!(out, "file path: {file_path_str}");