tustvold commented on code in PR #2044:
URL: https://github.com/apache/arrow-rs/pull/2044#discussion_r918234131
##########
parquet/src/file/serialized_reader.rs:
##########
@@ -560,12 +606,53 @@ impl<T: Read + Send> PageReader for
SerializedPageReader<T> {
Ok(None)
}
- fn peek_next_page(&self) -> Result<Option<PageMetadata>> {
- Err(nyi_err!("https://github.com/apache/arrow-rs/issues/1792";))
+ fn peek_next_page(&mut self) -> Result<Option<PageMetadata>> {
+ if let Some(page_offset_index) = &self.page_offset_index {
+ if self.seen_num_data_pages == page_offset_index.len() {
+ Ok(None)
+ } else if self.seen_num_data_pages == 0 &&
self.has_dictionary_page_to_read {
+ self.has_dictionary_page_to_read = false;
+ Ok(Some(PageMetadata {
+ num_rows: usize::MIN,
+ is_dict: true,
+ }))
+ } else {
+ let row_count = calculate_row_count(
+ page_offset_index,
+ self.seen_num_data_pages,
+ self.total_num_values,
+ )?;
+ Ok(Some(PageMetadata {
+ num_rows: row_count,
+ is_dict: false,
+ }))
+ }
+ } else {
+ Err(general_err!("Must set page_offset_index when using
peek_next_page in SerializedPageReader."))
+ }
}
fn skip_next_page(&mut self) -> Result<()> {
- Err(nyi_err!("https://github.com/apache/arrow-rs/issues/1792";))
+ if let Some(page_offset_index) = &self.page_offset_index {
+ let location = &page_offset_index[self.seen_num_data_pages];
+ let compressed_page_size = location.compressed_page_size;
+ //skip page bytes
+ let skip_size = io::copy(
+ self.buf.by_ref().take(compressed_page_size as u64).by_ref(),
Review Comment:
I think the best option here would be to store a list of Read within
SerializedPageReader, and have it work through them sequentially. This will
allow adding an additional constructor that would take the column index and the
ChunkReader, and create readers for each page up front.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
