1.
In the previous implementation, the max thread num of olap scanner was set relatively small, such as 3.
which would slow down some of queries.
In this PR, I changed the max thread num to a quarter of the scaner thread pool(default is 12),
which is less than the old scan node's max thread num, but larger than the previous implementation.
The upper limit of the max thread num of the old scan node is too high, which is not reasonable.
2.
Lower down the number of pre allocated free blocks.
1.remove quick_compaction's rowset pick policy, call cu compaction when trigger
quick compaction
2. skip tablet's compaction task when compaction score is too small
Co-authored-by: yixiutt <yixiu@selectdb.com>
1.
In the "explain" result of SQL, the table name in `ScanNode` should be full qualified with dbname.
And for olap scan node, the selected index name should not be "null".
2.
Remove `tpch_sf1_p1/tpch_sf1/nereids/` in regression test, it will be fixed later.
PR(https://github.com/apache/doris/pull/13404) introduced that ParquetReader
will break up batch insertion when encountering null values, which leads to the bad performance
compared to OrcReader.
So this PR has pushed null map into decode function, reduce the time of virtual function call
when encountering null values.
Further more, reuse hdfsFS among file readers to reduce the time of building connection to hdfs.
When execute shell command bash build.sh --be to build the backend, the cmake tool will show can't find the boost library, because the variable of BOOST_ROOT has some spelling mistake.
OS: Ubuntu 22.04 x86_64
CMake: 3.22.1
compiler: gcc (Ubuntu 11.2.0-7ubuntu2) 11.2.0
Previous logic of reverse function might not be strong enough to handle illegal character. For example, one one byte size character would be mistaken as one utf-8 character which occupies more than one byte space. And unfortunately exceeding the buffer space during future process.
1. this pr is used to fix the be core dump when select the invalid array.
2. before the change, we run "select array_intersect([1, 2, 3, 1, 2, 3], '1[3, 2, 5]');" will cause be core dump.
MySQL [example_db]> select array_intersect([1, 2, 3, 1, 2, 3], '1[3, 2, 5]');
ERROR 1105 (HY000): RpcException, msg: io.grpc.StatusRuntimeException: UNAVAILABLE: Network closed for unknown reason
3. after the change, we run "select array_intersect([1, 2, 3, 1, 2, 3], '1[3, 2, 5]');" will get error message.
MySQL [example_db]> select array_intersect([1, 2, 3, 1, 2, 3], '1[3, 2, 5]');
errCode = 2, detailMessage = No matching function with signature: array_intersect(array<tinyint(4)>, varchar(-1))"
Co-authored-by: hucheng01 <hucheng01@baidu.com>
Support delete from partitioned table without partition specified in [DELETE] stmt.
## Usage
If it is a partitioned table, you can specify a partition.
If not specified, Doris will infer partition from the given conditions.
In two cases, Doris cannot infer the partition from conditions:
1) the conditions do not contain partition columns;
2) The operator of the partition column is `not in`.
When a partition table does not specify the partition,
or the partition cannot be inferred from the conditions,
the session variable `delete_without_partition` needs to be `true`
to make delete statement be applied to all partitions.
## Test case
Test case is added in `regression-test/suites/delete_p0/test_delete_from_partition.groovy`,
user can delete from partitioned table without partition specified now.
There are some issues with the compile flag `-Wno-unused-but-set-variable` for clang.
1. `-Wno-unused-but-set-variable` should be set when building source by clang-15 on Linux. (#13000#13016)
2. On macOS Monterey, Apple Clang 13 may treat it as a unknown warning option and the compilation process may interrupt.
This PR introduces a better way to make this compile flag more portable.
1. Test whether the compiler recognizes this flag.
2. Add this flag if the compiler recognizes it.
When running regression test, we always found that the query return empty result after loading finished,
even if we call "sync" before the query.
This is because for `stream load`, the load task result will be returned immediately after the txn's status changed to VISIBLE,
but before writing the edit log.
So if we do the query right after we got the load task result, it is possible that we can not see the latest loaded data.
Same issue with `insert` operation
