database/mariadb-columnstore-engine
1
0
Fork 0
mirror of https://github.com/mariadb-corporation/mariadb-columnstore-engine.git synced 2025-08-05 16:15:50 +03:00
Code Activity
6,391 Commits 350 Branches 89 Tags
MCOL-5995-add-custom-bootstrap-params
Commit Graph

1222 Commits

Author SHA1 Message Date
Leonid Fedorov
cad6736d64 enum for SIMD out of ifdef 2022-02-11 18:18:23 +03:00
David Hall
27dea733c5 MCOL4841 dev port run large join without OOM 2022-02-09 17:33:55 -06:00
Roman Nozdrin
c79dfc4925 MCOL-4809 This patch adds support for float data types filtering and scanning vectorization 2022-02-03 16:38:56 +00:00
benthompson15
36775168d3 MCOL-4940: getLongDoubleVal was not handling all colDataType correctly (#2229) 2022-01-31 13:46:13 -06:00
Leonid Fedorov
04752ec546 clang format apply 2022-01-21 16:43:49 +00:00
Leonid Fedorov
6b6411229f build fixes 2022-01-21 16:34:04 +00:00
Leonid Fedorov
01f3ceb437 replace header guards with #pragma once 2022-01-21 15:24:58 +00:00
Marko Mäkelä
15da99477e MDEV-27519 CRC32() upon Columnstore table returns a wrong value 
Func_crc32::getIntVal(): Support the 2-ary CRC32() variant (MDEV-27208).
Also, do not assume that the string contains no NUL bytes.
 2022-01-21 09:35:19 +00:00
Denis Khalikov
1b6d278404 Handle error during parsing of bytestream. 
The error can occur in case we send the bytestream with `old` format header.
 2021-12-30 12:30:21 +03:00
Roman Nozdrin
af36f9940f This patch introduces support for scanning/filtering vectorized execution for numeric-based 
data types TEXT, CHAR, VARCHAR, FLOAT and DOUBLE are not yet supported by vectorized path
This patch introduces an example for Google benchmarking suite to measure a perf diff
b/w legacy scan/filtering code and the templated version
 2021-12-10 10:30:00 +00:00
Gagan Goel
99cad71f62 MCOL-4855 Do not log debug/info/warning messages to stderr. 2021-11-19 22:21:41 +00:00
Denis Khalikov
b382f681a1 [MCOL-4849] Parallelize the processing of the bytestream vector. 
This patch changes the logic of the `receiveMultiPrimitiveMessages`
function in the following way:

1. We have only one aggregation thread which reads the data from Queue (which is populated
by messages from BPPs).
2. Processing of the received `bytestream vector` could be in parallel depends on the
type of `TupleBPS` operation (join, fe2, ...) and actual thread pool workload.

The motivation is to eliminate some amount of context switches.
 2021-11-04 13:28:22 +03:00
Roman Nozdrin
61589c60d4 Merge pull request #2124 from denis0x0D/MCOL-4810_fix 
MCOL-4810 Add support for missed operation.
 2021-11-01 18:21:37 +03:00
Leonid Fedorov
c735c9fdc9 Fix build with duplicate lz4 header 2021-10-29 13:36:15 +00:00
Denis Khalikov
6393c6d019 MCOL-4810 Add support for missed operation for longStrings. 2021-10-28 10:02:02 +03:00
Roman Nozdrin
0324cedc88 This patch puts octet2hex symbol into a correct namespace 2021-10-25 18:08:18 +00:00
Roman Nozdrin
3de038c1da MCOL-4876 This patch enables continues buffer to be used by ColumnCommand and aligns BPP::blockData 
that in most cases was unaligned
 2021-10-06 09:23:40 +00:00
Roman Nozdrin
d2d2491921 Merge pull request #2065 from mariadb-AlexeyAntipovsky/MCOL-4829-dev 
[MCOL-4829] Compression for the temp disk-based aggregation files
 2021-09-13 23:13:53 +03:00
Vicențiu Ciorbaru
2556e783c6 Eliminate build race condition, messageids is generated 
This ensures that building loggingcpp has messageids.h generated
beforehand, just like errorids.h .
 2021-09-09 12:09:48 +03:00
Alexey Antipovsky
6a4140394d [MCOL-4829] More accurate memory counting 2021-09-07 19:52:20 +03:00
Alexey Antipovsky
7fea3c988e [MCOL-4829] Compression for the temp disk-based aggregation files 2021-09-02 19:30:25 +03:00
Roman Nozdrin
46cf13ffa8 Merge pull request #2101 from denis0x0D/MCOL-4810_2 
MCOL-4810 Redundant copying and wasting memory in PrimProc
 2021-08-27 14:05:51 +03:00
Denis Khalikov
7bda598fbf MCOL-4810 Redundant copying and wasting memory in PrimProc 
This patch eliminates a copying `long string`s into the bytestream.
 2021-08-26 12:16:23 +03:00
Leonid Fedorov
5c5f103f98 MCOL-4839: Fix clang build (#2100) 
* Fix clang build

* Extern C returned to plugin_instance

Co-authored-by: Leonid Fedorov <l.fedorov@mail.corp.ru>
 2021-08-23 10:45:10 -05:00
benthompson15
923bbf4033 MCOL-1356: Add convert_tz (#2099) 2021-08-19 17:47:10 -05:00
Leonid Fedorov
517e793843 One more bool* to bool cast bug (#2097) 2021-08-18 15:56:57 -05:00
Gagan Goel
98473a45cc Merge pull request #2079 from dhall-MariaDB/MCOL-3741 
Mcol 3741 Change IDB-xxxx error codes to MCS-xxxx
 2021-08-18 14:01:04 -04:00
Leonid Fedorov
dbb1269d69 GetInterrupted returned bool instead of bool * (#2085) 2021-08-18 11:50:45 -05:00
Leonid Fedorov
a2f441cd9c moda returned local object pointer (#2089) 2021-08-18 11:41:07 -05:00
Leonid Fedorov
3136e9dbab We forgot to initilize longdoublenull value (#2091) 2021-08-18 11:34:35 -05:00
Roman Nozdrin
150770b919 Merge pull request #2059 from mariadb-corporation/unittests-ctest 
Add ctest for google unittests
 2021-08-11 11:03:05 +03:00
David Hall
ecde2719b1 MCOL-3741 Change IDB-xxxx error codes to MCS-xxxx 2021-08-09 11:33:09 -05:00
David Hall
4d4dd22105 MCOL-4771 develop fix crash from rand() 2021-08-06 16:09:28 -05:00
Leonid Fedorov
73e710ed52 Add ctest for google unittests 2021-08-02 19:41:04 +03:00
Roman Nozdrin
4cdef40a55 Merge pull request #2052 from drrtuy/MCOL-4815 
MCOL-4815 ColumnCommand was replaced with a set of derived classes sp…
 2021-07-21 17:36:35 +03:00
Roman Nozdrin
a292585b8c MCOL-4815 ColumnCommand was replaced with a set of derived classes specified by 
column width

RTSCommand was modified to use a fabric that produces CC class based on column width

NB this patch doesn't affect PseudoCC that also leverages ColumnCommand
 2021-07-21 12:54:14 +00:00
Denis Khalikov
fa8dc815a7 MCOL-4814 Add a cmake build option to enable LZ4 compression. 
This patch adds an option for cmake flags to enable lz4 compression.
 2021-07-16 17:57:11 +03:00
Gagan Goel
3d557a2f1e Merge pull request #2044 from dhall-MariaDB/MCOL-3738 
MCOL-3738 COUNT(DISTINCT) with multiple parms
 2021-07-12 07:34:56 -04:00
Leonid Fedorov
51a8ffcb6a Fix sumavgoverflow.sql test 2021-07-09 22:41:28 +00:00
David Hall
76607be63a MCOL-3738 COUNT(DISTINCT) with multiple parms 
Fixed regression
Added a few more mtr tests
 2021-07-09 09:07:03 -05:00
Leonid Fedorov
f81f743282 Replace underlying type for avg and sum for int types from long double to wide decimal 2021-07-08 17:04:43 +00:00
David Hall
1113470551 MCOL-4738 AVG gives wrong results with strict_aliasing 
A f fix that works with strict_aliasing
 2021-07-07 13:08:32 -05:00
Roman Nozdrin
866dc25729 Merge pull request #1842 from denis0x0D/MCOL-987_LZ 
MCOL-987 LZ4 compression support.
 2021-07-07 13:13:18 +03:00
Roman Nozdrin
7b4f759592 Merge pull request #2032 from drrtuy/MCOL-4802 
MCOL-4802 Removed ByteStream methods for bool and add some logging in…
 2021-07-07 13:03:54 +03:00
Alexander Barkov
8988253ff4 Merge pull request #2031 from mariadb-corporation/bar-develop-MCOL-4801 
MCOL-4801 Replace Row methods getStringLength() and getStringPointer(…
 2021-07-07 13:53:19 +04:00
Roman Nozdrin
fb5ba84212 MCOL-4802 Removed ByteStream methods for bool manipulations and add some logging into I_S.columnstore_files 2021-07-07 07:16:30 +00:00
David Hall
8332ab8974 MCOL-4738 AVG() returns a wrong result 
On AMD64 machines, the fpu is 80 bits. The unused bits must be masked for memcmp to work properly. For other archetectures, we don't want to mask those bits.
 2021-07-06 19:50:00 -05:00
Alexander Barkov
9794f24369 MCOL-4801 Replace Row methods getStringLength() and getStringPointer() to getConstString() 2021-07-06 21:15:32 +04:00
Denis Khalikov
cc1c3629c5 MCOL-987 Add LZ4 compression. 
* Adds CompressInterfaceLZ4 which uses LZ4 API for compress/uncompress.
* Adds CMake machinery to search LZ4 on running host.
* All methods which use static data and do not modify any internal data - become `static`,
  so we can use them without creation of the specific object. This is possible, because
  the header specification has not been modified. We still use 2 sections in header, first
  one with file meta data, the second one with pointers for compressed chunks.
* Methods `compress`, `uncompress`, `maxCompressedSize`, `getUncompressedSize` - become
  pure virtual, so we can override them for the other compression algos.
* Adds method `getChunkMagicNumber`, so we can verify chunk magic number
  for each compression algo.
* Renames "s/IDBCompressInterface/CompressInterface/g" according to requirement.
 2021-07-06 18:04:37 +03:00
Gagan Goel
8520f87237 MCOL-641 Cleanup. 2021-07-06 09:01:49 +00:00