You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by GitBox <gi...@apache.org> on 2022/09/13 13:51:34 UTC
[GitHub] [doris] AshinGau opened a new pull request, #12570: [feature-wip](parquet-reader) create file reader for each column to avoid time consuming of hdfsSeek
AshinGau opened a new pull request, #12570:
URL: https://github.com/apache/doris/pull/12570
# Proposed changes
Issue Number: close #xxx
## Problem summary
Describe your changes.
## Checklist(Required)
1. Does it affect the original behavior:
- [ ] Yes
- [x] No
- [ ] I don't know
2. Has unit tests been added:
- [ ] Yes
- [x] No
- [ ] No Need
3. Has document been added or modified:
- [ ] Yes
- [x] No
- [ ] No Need
4. Does it need to update dependencies:
- [ ] Yes
- [x] No
5. Are there any changes that cannot be rolled back:
- [ ] Yes (If Yes, please explain WHY)
- [x] No
## Further comments
If this is a relatively large or complex change, kick off the discussion at [dev@doris.apache.org](mailto:dev@doris.apache.org) by explaining why you chose the solution you did and what alternatives you considered, etc...
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [doris] morningman closed pull request #12570: [feature-wip](parquet-reader) create file reader for each column to avoid time consuming of hdfsSeek
Posted by GitBox <gi...@apache.org>.
morningman closed pull request #12570: [feature-wip](parquet-reader) create file reader for each column to avoid time consuming of hdfsSeek
URL: https://github.com/apache/doris/pull/12570
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [doris] morningman commented on a diff in pull request #12570: [feature-wip](parquet-reader) create file reader for each column to avoid time consuming of hdfsSeek
Posted by GitBox <gi...@apache.org>.
morningman commented on code in PR #12570:
URL: https://github.com/apache/doris/pull/12570#discussion_r969797638
##########
be/src/io/hdfs_file_reader.cpp:
##########
@@ -147,6 +147,7 @@ Status HdfsFileReader::readat(int64_t position, int64_t nbytes, int64_t* bytes_r
BackendOptions::get_localhost(), _namenode, _path,
hdfsGetLastError());
}
+ _current_offset = position;
Review Comment:
Directly call `HdfsFileReader::seek() here`, to remove duplicate code
##########
be/src/vec/exec/format/parquet/vparquet_page_reader.cpp:
##########
@@ -38,7 +36,7 @@ Status PageReader::next_page_header() {
const uint8_t* page_header_buf = nullptr;
size_t max_size = _end_offset - _offset;
- size_t header_size = std::min(initPageHeaderSize, max_size);
+ size_t header_size = 128;
Review Comment:
Do not use this magic number.
why remove `initPageHeaderSize`?
##########
be/src/vec/exec/format/parquet/vparquet_group_reader.cpp:
##########
@@ -50,7 +50,20 @@ Status RowGroupReader::_init_column_readers(const FieldDescriptor& schema,
TypeDescriptor col_type = slot_desc->type();
auto field = const_cast<FieldSchema*>(schema.get_column(slot_desc->col_name()));
std::unique_ptr<ParquetColumnReader> reader;
- RETURN_IF_ERROR(ParquetColumnReader::create(_file_reader, field, read_col, _row_group_meta,
+ FileReader* buff_reader;
+ if (_buffered_file_reader.size() < config::parquet_group_pooled_reader) {
Review Comment:
I think it is too many readers for a single row group.
##########
be/src/vec/exec/format/parquet/vparquet_reader.h:
##########
@@ -130,6 +128,7 @@ class ParquetReader {
int64_t _range_size;
cctz::time_zone* _ctz;
std::vector<RowRange> _skipped_row_ranges;
+ std::unique_ptr<BufferedReader> _file_reader;
Review Comment:
Is this reader only for reading meta?
--
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: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org