in aggregation code
The patch disables padding that forces hasher to calculate over the whole 2k buffer. This patch also moves hashing code
into the common place where it belongs.
When length of string to replace minus length of string to replace to is
bigger than input string and processing mode allows for binary (memcmp
or std::string::find()) comparison, REPLACE may trigger invalid capacity
assertion and query processing will stop.
The fix is to properly count the number of occurences of the string to
replace, basically.
We have added a new function, SystemCatalog::upgrade(), which will
create a new column in the database and fill it with default values.
It uses a reference column to fill the column, which is passed as a
std::unordered_map<key, value> value to SystemCatalog::upgrade()
function, where key is the target column which needs to be created.
In addition, we pass additional std::unordered_map's to
SystemCatalog::upgrade() containing information such as the datatypes
and colwidths of the target and reference columns, as well as the
default value which will be used to fill the target column.
* more build dependencies
* fix for cmake < 3.11
It cannot do ADD_LIBRARY(... ALIAS ...) on IMPORTED targets
* another fix for cmake 3.10.2
It doesn't know about CMAKE_CXX_STANDARD=20,
let's add the correct flag manually
* gcc 8 on aarch64
utils/common/simd_arm.h:241:16: error: need ‘typename’ before ‘simd::TypeToVecWrapperType<T>::WrapperType’ because ‘simd::TypeToVecWrapperType<T>’ is a dependent scope
* build boost during build phase, not during configure
* add dependency for generated header files errorids.h messageids.h
see 7e868bc588
* set explicit dependencies on external_boost for #include's
* clang-14 compatibility fix
* MCOL-5092 Ensure column width is correct for datatype
Change MODA return type to STRING
Modify MODA to handle every numeric type
* MCOL-5162 MODA to support char and varchar with collation support
Fixes to the aggregate bit functions
When we fixed the storage sign issue for MCOL-5092, it uncovered a problem in the bit aggregates (bit_and, bit_or and bit_xor). These aggregates should always return UBIGINT, but they relied on the type of the argument column, which gave bad results.
* Add check for simd acrh support
* Updates
* More polite and detailed error messages
* Updates
* Always true to conditional
Co-authored-by: Leonid Fedorov <leonid.fedorov@mariadb.com>
This patch adds support for on clause filter for a table which is not involved in particular join
by disabling an `merge optimization` for those particular cases.
The `merge optimization` is optimization when CS
tries to create a one BPP join with one `large side` table and multiple `small sides` tables, in this
case we cannot apply a FE filter if this filter requires a columns from `small side` table which is not
involved in particular join.
* Restructured test suites and added autopilot and extended suites
* Updated autopilot with correct branch - develop
* Moved setup test case to a 'setup' directory, for consistency
* Fixed a path issue
* Updated some tests cases to keep up with development
Co-authored-by: root <root@rocky8.localdomain>
exact functionality that does not use MDB hash function.
This patch also takes a bit from Robin Hood hash map implementation forgotten
that reduces hash function collision rate.
