See #17764 for details
I have tested:
- Unit test for local/s3/hdfs/broker file system: be/test/io/fs/file_system_test.cpp
- Outfile to local/s3/hdfs/broker.
- Load from local/s3/hdfs/broker.
- Query file on local/s3/hdfs/broker file system, with table value function and catalog.
- Backup/Restore with local/s3/hdfs/broker file system
Not test:
- cold & host data separation case.
There are many type definitions in BE. Should unify the type system and simplify the development.
---------
Co-authored-by: yiguolei <yiguolei@gmail.com>
Before this PR when encountering null values with some columns which is specified as `NOT NULL`, null values will not be filtered,thi behavior does not match with the original load behavior.
Second column alignment logic has bug :
```
template <typename ColumnInserterFn>
void align_variant_by_name_and_type(ColumnObject& dst, const ColumnObject& src, size_t row_cnt,
ColumnInserterFn inserter) {
CHECK(dst.is_finalized() && src.is_finalized());
// Use rows() here instead of size(), since size() will check_consistency
// but we could not check_consistency since num_rows will be upgraded even
// if src and dst is empty, we just increase the num_rows of dst and fill
// num_rows of default values when meet new data
size_t num_rows = dst.rows();
```
basic functions for map datatype:
- MAP<K, V> map(K k1, V v1, ...)
- BIGINT map_size(MAP<K, V> m)
- BOOL map_contains_key(MAP<K, V> m, K k1)
- BOOL map_contains_value(MAP<K, V> m, V v1)
- ARRAY< K> map_keys(MAP<K, V> m)
- ARRAY< V> map_values(MAP<K, V> m)
When compaction case, memory map offsets coming to same olap convertor which is from 0 to 0+size
but it should be continue in different pages when in one segment writer .
eg :
last block with map offset : [3, 6, 8, ... 100]
this block with map offset : [5, 10, 15 ..., 100]
the same convertor should record last offset to make later coming offset followed last offset.
so after convertor :
the current offset should [105, 110, 115, ... 200], then column writer just call append_data() to make the right offset data append pages
* [Feature](vectorized)(quantile_state): support vectorized quantile state functions
1. now quantile column only support not nullable
2. add up some regression test cases
3. set default enable_quantile_state_type = true
---------
Co-authored-by: spaces-x <weixiang06@meituan.com>
1. introduce a new type `VARIANT` to encapsulate dynamic generated columns for hidding the detail of types and names of newly generated columns
2. introduce a new expression `SchemaChangeExpr` for doing schema change for extensibility
Support delta encoding and rle(bool) to read Glue data
add delta bit pack decoder,
add delta length byte array decoder,
add delta byte array decoder.
add rle bool decoder.
We find some data type is read with delta encoding on AWS Glue, so it should be supported.
The definition of delta encoding can refer to the delta encoding in parquet.
1. add support `CAST AS Struct` from Struct type;
2. fix crash while `CAST('{}' AS Struct)`;
3. `CAST('' AS complext_type)` should return NULL instead of empty object;
---------
Co-authored-by: cambyzju <zhuxiaoli01@baidu.com>
the input replicate_offsets should be the same size as ColumnArray's offset.
```
IColumn::Offsets replicate_offsets(get_offsets().size(), 0);
// |---------------------|-------------------------|-------------------------|
// [0, begin) [begin, begin + count_sz) [begin + count_sz, size())
// do not need to copy copy counts[n] times do not need to copy
```
we should
The current distribution model for Doris is as follows:
OlapTableSink seperate the original Block into serveral subblocks of each node(BE) by tablets distribution and distributes subblocks to storage engine of backends, then the storage engine will seperate the subblock into multiple tablets channel and each delta writer will handle partial of the block.
This model causes blocks to be split according to tablets, and the splitting process can be a relatively heavy operation. After splitting, the blocks are distributed to different DeltaWriters (Memtables) through RPCs to TabletChannels. The distribution operation on TabletChannels is also a relatively heavy operation. If the distribution property of the table is RANDOM distribution, then we have the opportunity to distribute the blocks according to the complete block during distribution. The advantage of doing so is to reduce memory copying and improve write locality, similar to appending the entire block to the memtable.
This optimze could save 10% ~ 20% CPU cost of RANDOM distribution table load when enable load_to_single_tablet
In the past, only simple predicates (slot=const), and, like, or (only bitmap index) could be pushed down to the storage layer. scan process:
Read part of the column first, and calculate the row ids with a simple push-down predicate.
Use row ids to read the remaining columns and pass them to the scanner, and the scanner filters the remaining predicates.
This pr will also push-down the remaining predicates (functions, nested predicates...) in the scanner to the storage layer for filtering. scan process:
Read part of the column first, and use the push-down simple predicate to calculate the row ids, (same as above)
Use row ids to read the columns needed for the remaining predicates, and use the pushed-down remaining predicates to reduce the number of row ids again.
Use row ids to read the remaining columns and pass them to the scanner.
