[POC] Metadata index for Parquet files

parquet/src/arrow/mod.rs

@@ -478,7 +478,7 @@ mod test {
             .unwrap();
         assert_eq!(
             err.to_string(),
-            "EOF: Parquet file too small. Page index range 82..115 overlaps with file metadata 0..357"
+            "EOF: Parquet file too small. Page index range 82..115 overlaps with file metadata 0..392"
         );
     }
 

parquet/src/bloom_filter/mod.rs

@@ -284,10 +284,12 @@ impl Sbbf {
     /// flush the writer in order to boost performance of bulk writing all blocks. Caller
     /// must remember to flush the writer.
     pub(crate) fn write<W: Write>(&self, mut writer: W) -> Result<(), ParquetError> {
-        let mut protocol = ThriftCompactOutputProtocol::new(&mut writer);
-        self.header().write_thrift(&mut protocol).map_err(|e| {
-            ParquetError::General(format!("Could not write bloom filter header: {e}"))
-        })?;
+        {
+            let mut protocol = ThriftCompactOutputProtocol::new(&mut writer);
+            self.header().write_thrift(&mut protocol).map_err(|e| {
+                ParquetError::General(format!("Could not write bloom filter header: {e}"))
+            })?;
+        }
         self.write_bitset(&mut writer)?;
         Ok(())
     }

parquet/src/file/metadata/mod.rs

@@ -1628,8 +1628,10 @@ mod tests {
             .unwrap();
 
         let mut buf = Vec::new();
-        let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
-        row_group_meta.write_thrift(&mut writer).unwrap();
+        {
+            let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
+            row_group_meta.write_thrift(&mut writer).unwrap();
+        }
 
         let row_group_res = read_row_group(&mut buf, schema_descr).unwrap();
 
@@ -1710,8 +1712,10 @@ mod tests {
             .build()
             .unwrap();
         let mut buf = Vec::new();
-        let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
-        row_group_meta_2cols.write_thrift(&mut writer).unwrap();
+        {
+            let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
+            row_group_meta_2cols.write_thrift(&mut writer).unwrap();
+        }
 
         let err = read_row_group(&mut buf, schema_descr_3cols)
             .unwrap_err()
@@ -1761,8 +1765,10 @@ mod tests {
             .unwrap();
 
         let mut buf = Vec::new();
-        let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
-        col_metadata.write_thrift(&mut writer).unwrap();
+        {
+            let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
+            col_metadata.write_thrift(&mut writer).unwrap();
+        }
         let col_chunk_res = read_column_chunk(&mut buf, column_descr).unwrap();
 
         assert_eq!(col_chunk_res, col_metadata);
@@ -1777,8 +1783,10 @@ mod tests {
             .unwrap();
 
         let mut buf = Vec::new();
-        let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
-        col_metadata.write_thrift(&mut writer).unwrap();
+        {
+            let mut writer = ThriftCompactOutputProtocol::new(&mut buf);
+            col_metadata.write_thrift(&mut writer).unwrap();
+        }
         let col_chunk_res = read_column_chunk(&mut buf, column_descr).unwrap();
 
         assert_eq!(col_chunk_res, col_metadata);