Commit Graph

115 Commits

Author SHA1 Message Date
Azat Khuzhin
88bfb788a9 Add ability to use multiple disks/volumes for temporary data
This patch adds <tmp_policy> config directive, that will define the
policy to use for storing temporary files, if it is not set (default)
the <tmp_path> will be used.

Also tmp_policy has some limitations:
- move_factor              is ignored
- keep_free_space_bytes    is ignored
- max_data_part_size_bytes is ignored
- must have exactly one volume
2020-01-23 20:31:01 +03:00
Nikolai Kochetov
c2a054ac70 Add cancell flag to MergingAggrgated. 2020-01-13 12:56:21 +03:00
Amos Bird
3707da4fbf
Better addBatchArray aggregator 2019-11-11 16:56:11 +08:00
alexey-milovidov
1d910c5071
Merge pull request #7608 from amosbird/batchreduce
Optimize arrayReduce, -Array and -State combinators
2019-11-10 22:10:03 +03:00
Amos Bird
544af39d23
Vectorize arrayReduce and -Array combinator
Also devirtualize -State combinator.
2019-11-07 11:53:47 +08:00
Amos Bird
fbfbe161ec
Unify hash tables interface. 2019-11-06 00:08:04 +08:00
Amos Bird
a864447802 Introduce String Hash Map.
It speeds up aggregation over short string keys. Use it as a default
aggregation method for string keys.
2019-10-21 16:27:35 +03:00
chertus
c46cc3ff1a refactoring: remove copy-paste 2019-10-01 21:51:33 +03:00
Nikolai Kochetov
1319168a3a Merge branch 'master' into processors-2 2019-09-12 16:28:51 +03:00
Amos Bird
8f28a36c17 Use internal iteration over hash tables in Aggregator.
For compound hash tables such as the future StringHashMap, an
iterator-based API might be inefficient for iterating over a table or
for merging two tables, because:

1) the key has to be converted to a general format from a from a
   component-specific format, which may differ between the components;

2) the information about the component of the compound hash table to
   which the value belongs is lost, and has to be recalculated if the
   value is reinserted.

A more efficient approach is to use internal iteration, that is,
map-like functions, which avoids unnecessary conversions when iterating,
and allows to use an efficient component-wise approach when merging.
2019-09-10 21:52:53 +03:00
Alexander Kuzmenkov
160e0faf77 Do not use iterator::getValue() in Aggregator.
Use separate key and "mapped" value references instead. This is
important for hash tables that do not store the key/"mapped" pair
directly, and cannot provide this interface without some runtime
overhead.
2019-09-10 20:12:30 +03:00
Nikolai Kochetov
52305cefbc Merge branch 'merging-aggregated-processors' into processors-2 2019-09-06 15:21:17 +03:00
Nikolai Kochetov
d10b7d7a2b Added ConvertingAggregatedToChunksTransform. 2019-09-06 15:19:59 +03:00
Nikolai Kochetov
bf0c0ed4f1 Merge branch 'master' into processors-2 2019-09-02 16:02:09 +03:00
Nikolai Kochetov
300dc80194 Update Aggregator interface. 2019-08-31 11:58:16 +03:00
alexey-milovidov
e2b05f6f9c
Merge pull request #6700 from yandex/aku/shm-5417-formatting
Formatting changes for StringHashMap PR #5417.
2019-08-28 18:30:59 +03:00
Alexander Kuzmenkov
0b5a7f5f4a Disable consecutive key optimization for UInt8/16 LowCardinality
columns.

A follow-up to #6298, this is as well a part of StringHashMap PR #5417
by Amos Bird.
2019-08-28 14:05:52 +03:00
Alexander Kuzmenkov
24a8755b82 Formatting changes for StringHashMap PR #5417. 2019-08-28 13:52:14 +03:00
Weiqing Xu
2fbcd607f3 check free space when use external sort/aggerator 2019-08-27 08:23:07 +08:00
Alexey Milovidov
cdd6dca517 Remove Compiler 2019-08-24 00:10:26 +03:00
Alexey Milovidov
8966520801 Added "assert_cast" function as a replacement for static_cast for columns 2019-08-21 05:28:04 +03:00
Alexey Milovidov
2570907f44 Removed useless parameter 2019-08-12 00:45:18 +03:00
Alexey Milovidov
011e50cae6 Removed useless function arguments 2019-08-11 02:27:25 +03:00
Alexey Milovidov
efa51a6cd9 Batch aggregation (experimental) 2019-08-11 01:36:55 +03:00
alexey-milovidov
349ad17436
Merge pull request #6298 from yandex/aku/uint816-nocache
Disable consecutive key optimization for UInt8/16.
2019-08-03 06:23:03 +03:00
Alexander Kuzmenkov
d808fafa8f Disable consecutive key optimization for UInt8/16.
These types use a FixedHashMap for aggregation, which makes lookup
almost free, so we don't have to cache the last lookup result.

This is a part of StringHashMap PR #5417 by Amos Bird.
2019-08-02 18:37:32 +03:00
Alexander Kuzmenkov
011150fa8a Make PairNoInit a simple struct. 2019-08-01 19:54:33 +03:00
Max Akhmedov
5e9ea2a905
Remove spaces from empty line 2019-07-28 20:46:48 +03:00
Max Akhmedov
f632a0b241
Fix build under gcc-8.2 2019-07-28 18:12:07 +03:00
Nikolai Kochetov
209793c981 Merged with master. 2019-07-08 16:00:54 +03:00
chertus
8932dc9c23 minpr perf improvement 2019-07-02 15:31:20 +03:00
chertus
0c634129da add IColumn.isNullable() 2019-07-01 14:44:19 +03:00
chertus
db7c8748e9 remove IColumn.isColumnNullable() 2019-06-27 21:50:20 +03:00
chertus
65c0d23efa better casts to ColumnNullable 2019-06-26 20:20:33 +03:00
Nikolai Kochetov
f9b29bfdd2 Merged with master. 2019-06-25 20:00:54 +03:00
Ivan Lezhankin
0821d62516 Merge ITableDeclaration into IStorage.
Also a couple of massive include-refactorings with better forwarding
2019-05-17 17:34:25 +03:00
Nikolai Kochetov
808d0f3b08 Merged with master. 2019-05-14 13:31:19 +03:00
Alexey Milovidov
58d7d1e451 Resolved some of PVS-Studio warnings 2019-04-22 19:07:09 +03:00
Nikolai Kochetov
5e39ec610c Merged with master 2019-04-03 14:21:38 +03:00
Nikolai Kochetov
6976b65b2d Added MergingAggregatedTransform. 2019-04-03 13:13:25 +03:00
Amos Bird
26ab5dd7a7 A Proper lookup table that uses HashTable's API
This is the first step of allowing heterogeneous cells in hash tables.

performance test results are

```

1. HashMap<UInt16, UInt8, TrivialHash, HashTableFixedGrower<16>>;
2. NewLookupMap<UInt16, UInt8>

ResolutionWidth 30000 1 .................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................223550276.46
ResolutionWidth 30000 2 .................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................248772721.24
Best: 2 - 24877272124

ResolutionWidth 100000 1 ..........................................................................................................................................................................................................................................................238498413.99
ResolutionWidth 100000 2 ..........................................................................................................................................................................................................................................................261808889.98
Best: 2 - 26180888998

ResolutionWidth 300000 1 ...................................................................................239307348.81
ResolutionWidth 300000 2 ...................................................................................257592761.30
Best: 2 - 25759276130

ResolutionWidth 1000000 1 .........................240144759.26
ResolutionWidth 1000000 2 .........................257093531.91
Best: 2 - 25709353191

ResolutionWidth 5000000 1 .....241573260.35
ResolutionWidth 5000000 2 .....259314162.79
Best: 2 - 25931416279

ResolutionDepth 30000 1 .................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................217108119.84
ResolutionDepth 30000 2 .................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................249459504.41
Best: 2 - 24945950441

ResolutionDepth 100000 1 ..........................................................................................................................................................................................................................................................229065162.17
ResolutionDepth 100000 2 ..........................................................................................................................................................................................................................................................253769105.64
Best: 2 - 25376910564

ResolutionDepth 300000 1 ...................................................................................233079225.18
ResolutionDepth 300000 2 ...................................................................................256316273.78
Best: 2 - 25631627378

ResolutionDepth 1000000 1 .........................234184633.51
ResolutionDepth 1000000 2 .........................261100491.57
Best: 2 - 26110049157

ResolutionDepth 5000000 1 .....233118795.66
ResolutionDepth 5000000 2 .....252436160.41
Best: 2 - 25243616041

```
2019-03-01 16:47:13 +08:00
Nikolai Kochetov
cf40a1538b Fix size() and empty() for AggregationDataWithNullKey. #4223 2019-02-13 19:54:55 +03:00
Nikolai Kochetov
2c2932e185 Merged with master 2019-01-29 15:25:19 +03:00
Nikolai Kochetov
6fce028b56 Refactor ColumnsHashing. 2019-01-21 13:39:53 +03:00
Nikolai Kochetov
d207498573 Added ColumnsHashing 2019-01-21 13:39:24 +03:00
Alexey Milovidov
abcd5a2a49 Attempt to implemnt global thread pool #4018 2019-01-11 22:12:36 +03:00
Alexey Milovidov
06a8c3f02d Added some warnings from clang's -Weverything 2019-01-05 06:33:22 +03:00
Alexey Milovidov
1d4701b0bc Added some warnings from clang's -Weverything 2019-01-04 15:10:00 +03:00
Alexey Milovidov
e33e5150b7 Miscellaneous [#CLICKHOUSE-2] 2019-01-02 09:44:36 +03:00
Alexey Milovidov
b109ec8c77 Fixed UB #3920 2018-12-27 23:24:37 +03:00