Continuous Benchmarking

SQL Code

COPY lineitem
    'lineitem_unicode.csv' DELIMITER '🦆' HEADER
c26b2020-02-250.93[L/E]Merge pull request #433 from cwida/idx_t2 Rename index_t to idx_t because Solaris
b0782020-02-250.93[L/E]Merge remote-tracking branch 'origin/master' into idx_t2
4c082020-02-250.93[Q/L/E]Fix for quoted range var without schema in Catalog::ParseRangeVar + tests
62a52020-02-250.92[Q/L/E]Unify code to parse schema/table from sequences with R client appender, and add correct support for quotes to it. This fixes #430
31b42020-02-250.92[Q/L/E]Merge pull request #431 from cwida/stringrep Rework String Representation
dbd92020-02-240.92[Q/L/E]Fixes for R client and rest server
2a552020-02-240.92[Q/L/E]Revert "Remove string inlining (for benchmark reasons)" This reverts commit ca74946959d0449dc781ac23b910180a5201dcd8.
ca742020-02-240.98[Q/L/E]Remove string inlining (for benchmark reasons)
5b232020-02-240.93[Q/L/E]Merge branch 'master' into stringrep
4ee72020-02-240.92[Q/L/E]Add new set of micro benchmarks involving string operations
c13c2020-02-200.94[Q/L/E]Merge pull request #427 from Honeypot95/reverse_function2 Implement the reverse function.
15ea2020-02-200.94[Q/L/E]Merge pull request #426 from cwida/changevectorsize Fixes for support of different vector sizes
7d602020-02-200.94[Q/L/E]Test even fewer vector sizes to limit testing time and add extra string agg test case
bb292020-02-200.93[Q/L/E]Limit the amount of vector sizes tested to avoid the time limit on travis
bfa42020-02-200.94[Q/L/E]Remove -j from build for travis
7d3f2020-02-200.94[Q/L/E]Travis: try not explicitly setting the compiler
13d52020-02-190.94[Q/L/E]Add ubuntu toolchain to travis build
1a322020-02-190.93[Q/L/E]Add travis test for different vector sizes
52a52020-02-190.93[Q/L/E]Allow setting of vector size through cmake and add python script for testing different vector sizes
51982020-02-190.96[Q/L/E]Fix explain for VECTOR_SIZE=2
fa2d2020-02-190.95[Q/L/E]Vector_size=4 working
b92b2020-02-190.94[Q/L/E]Fixes for vector_size=8; fix for local storage when all values in a vector are deleted
d6602020-02-190.94[Q/L/E]Add test that throws an exception in the middle of an aggregate with destructor being constructed
d3022020-02-190.94[Q/L/E]Add destructor function for aggregates, and use it in STRING_AGG function to properly manage allocation of string aggregates
65622020-02-190.94[Q/L/E]Allow VECTOR_SIZE=16, fix off-by-one in assertion and fix for NLJ with NULL values
6ee22020-02-190.93[Q/L/E]Allow VECTOR_SIZE=32
ff852020-02-190.92[Q/L/E]Fixes for vector_size=64, including a fix to NLJ and a fix to properly set vector types in Gather::Set
ab062020-02-190.94[Q/L/E]Fixes for vector_size = 2048
4d5d2020-02-190.92[Q/L/E]Fixes for VECTOR_SIZE=512
8ca42020-02-180.94[Q/L/E]Add test for loading a non-database file in the shell-test
46bd2020-02-180.94[Q/L/E]Allow nullptr to be passed to sqlite3_open
fb3f2020-02-180.96[Q/L/E]Fix for printing profiling info related to prepared statements (this fixes profile info in shell)
edd32020-02-180.93[Q/L/E]Avoid crash in shell when opened with a non-db file
afbf2020-02-180.93[Q/L/E]Fix for NULL values in nextval function
89302020-02-180.94[Q/L/E]Merge pull request #423 from cwida/removedupinfofromvector Remove Individual Vector Cardinalities
70dc2020-02-180.95[Q/L/E]Fix assert restrict
19bc2020-02-170.94[Q/L/E]ASSERT_RESTRICT: report file and line number, and change to inclusive bounds
42d02020-02-170.94[Q/L/E]Fixes for embedded C and R package
23f12020-02-170.99[Q/L/E]Merge branch 'master' into removedupinfofromvector
a5532020-02-170.94[Q/L/E]Fix incorrect cardinality setting in join cache
9e362020-02-170.94[Q/L/E]Fix for JoinOrderOptimizer with cross products
b4932020-02-170.94[Q/L/E]Fix for large aggregate hashtables
9ef72020-02-170.94[Q/L/E]Fix several other issues with setting incorrect cardinality
8bb62020-02-140.00[Q/L/E]Fix several issues with correctly setting cardinality, more fixes required
7aa42020-02-130.96[Q/L/E]WindowSegmentTree and Vector are no longer friends: now DataChunk is Vectors best friend
bca82020-02-130.95[Q/L/E]Remove DataChunk::owned_data, and now initialize data chunks normally
91b32020-02-120.95[Q/L/E]Remove unused DataChunk::Move method
9ab72020-02-120.96[Q/L/E]Rename Flatten to ClearSelectionVector
f8ed2020-02-120.94[Q/L/E]Merge pull request #421 from cwida/bsd OpenBSD build changes from @videlft
920b2020-02-110.93[Q/L/E]Merge pull request #419 from cwida/multilineshell Fixing multiline query pasting in shell
1b0c2020-02-110.93[Q/L/E]Merge pull request #418 from samansmink/date-truncate-function Date truncate function
0bbf2020-02-110.93[Q/L/E]refactor date_trunc to use date_part_specifier
a7042020-02-110.94[Q/L/E]fixed wrongly reformatted template
43802020-02-110.92[Q/L/E]fixed typo in variable name
966d2020-02-110.92[Q/L/E]ran make format on new date_trunc function code
ba2c2020-02-110.92[Q/L/E]added date_trunc function for timestamp and date types. interval type not yet supported.
fcc22020-02-100.94[Q/L/E]Merge pull request #416 from cwida/tablestar Support for select table.* from table, issue #415
1c762020-02-100.93[Q/L/E]support for select table.* from table, issue #415
e10d2020-02-080.95[Q/L/E]First part of nested types: STRUCTS (#414) We add a new type `STRUCT` that can hold a named ordered list of other types. For now, two functions are supported: `STRUCT_PACK()` to create structs, for example `STRUCT_PACK(a := 42, b := 43)` to create a `STRUCT` of two integers named `a` and `b`. Then, there is `STRUCT_EXTRACT()` that returns one of the entries of a struct again. For example, `SELECT STRUCT_EXTRACT(STRUCT_PACK(a := 42, b := 43), 'a')` returns `42`.
67942020-02-050.93[Q/L/E]Fix error message
73c22020-02-050.93[Q/L/E]Merge branch 'master' of
ce8c2020-02-050.93[Q/L/E]Disallow ORDER and FILTER for aggregates (#408)
874e2020-02-050.93[Q/L/E]Remove unused in-place binary operators
a5a42020-02-050.92[Q/L/E]Using `Depends` for DBI in R package (#411) * using depends for DBI
bd052020-02-040.93[Q/L/E]Enable ccache in CMakeList
3cfb2020-02-040.92[Q/L/E]Move NotNullSelVector to VectorOperations
2fbd2020-02-040.93[Q/L/E]Remove Vector::Move and replace its usage with Vector::Reference
1ded2020-02-040.93[Q/L/E]Disallow DISTINCT in Window Functions (Fixes #408)
fa482020-02-040.93[Q/L/E]Merge pull request #409 from cwida/vectortypes Vector Rework (Vector types, three buffer system, arrow compatible TypeIds)
90ad2020-02-040.92[Q/L/E]Merge branch 'master' into vectortypes
33912020-02-040.92[Q/L/E]Merge pull request #404 from kryonix/master Support for Recursive Common Table Expressions
bdad2020-02-030.93[Q/L/E]version bump r package
59f82020-01-310.93[Q/L/E]Merge pull request #406 from cwida/adaptive-reordering Adaptive Expression Reordering
643b2020-01-270.92[Q/L/E]Added q1 benchmark with int group keys
c44f2020-01-250.92[Q/L/E]Merge pull request #402 from cwida/artstring Support for String/Compound Keys
f3372020-01-240.91[Q/L/E]Adding myth tests
faea2020-01-240.93[Q/L/E]Throwing exception when trying to create index on columns with duplicate data
00ed2020-01-230.92[Q/L/E]Indexes for temporary tables
fd2b2020-01-220.92[Q/L/E]Merge branch 'master' into artstring
4c8b2020-01-220.92[Q/L/E]Adding test on uniquness and prefix changes on enough data for all ART nodes
1efb2020-01-220.93[Q/L/E]Fixing issue on prefix copy
70c22020-01-210.92[Q/L/E]Have Travis generate the R repo metadata files `PACKAGES` and upload to DuckDB Download server
7b312020-01-200.93[Q/L/E]Using linenoise instead of readline for shell (#393) Also fixes a windows bug
7e852020-01-190.92[Q/L/E]Switching to stock SQLite shell (#388) We now build a statically linked command line interface `duckdb_cli` built from the stock SQLite shell.c sources.
2deb2020-01-140.93[Q/L/E]sub-second precision for timestamps in python api (#391) * sub-second precision for timestamps in python API
57b22020-01-110.93[Q/L/E]Merge pull request #390 from cwida/unusedcolumns Add Unused Column & Column Lifetime Optimizers
e1942020-01-110.94[Q/L/E]Merge branch 'master' into unusedcolumns
a7232020-01-110.94[Q/L/E]Fix for unused column join test and add additional test
5d2b2020-01-110.95[Q/L/E]Perform IN clause optimizer before join ordering/unused column elimination/column lifetime optimizer
77fd2020-01-110.92[Q/L/E]Add extra tests for joins with unused columns
b8862020-01-110.92[Q/L/E]Add benchmarks for Q1 aggregate after join, one with a filter and one without a filter
17242020-01-110.93[Q/L/E]Properly reset build chunk before filling it in hash join
4b752020-01-110.92[Q/L/E]Add separate ColumnLifetimeOptimizer that removes columns after joins/filters when they are no longer required
489b2020-01-100.93[Q/L/E]Remove filter mapping from RemoveUnusedColumns optimizer
1bfa2020-01-100.92[Q/L/E]When we have a filter with a projection map, instead of performing the projection within the filter we simply generate an extra projection to perform the projection
f1222020-01-100.94[Q/L/E]In the case of an equality inner join, remap any references of the RHS column to the LHS column (e.g. if there is l_orderkey=o_orderkey, anything that refers to o_orderkey will be remapped to instead refer to l_orderkey, this prevents us from having to read the o_orderkey column from the hashtable
087d2020-01-080.93[Q/L/E]Merge pull request #387 from informagi/master Include default instructions
4cc72020-01-080.93[Q/L/E]Include default instructions
33212020-01-070.93[Q/L/E]Merge pull request #381 from informagi/master Minor change for local install
c4062020-01-070.99[Q/L/E]Minor edit of PYTHONPATH instruction.
0e4e2020-01-050.93[Q/L/E]Changes to complete the suggested better solution by @xhochy.
7ed52020-01-050.93[Q/L/E]Update tools/pythonpkg/ Excellent suggestion by @xhochy Co-Authored-By: Uwe L. Korn
6fcb2020-01-030.93[Q/L/E]Merge pull request #383 from cwida/scan Join Hash Table Rework
31562020-01-030.92[Q/L/E]Revert join order cost change
d4462020-01-030.91[Q/L/E]Another ORDER BY required after hash function change
1d8c2020-01-030.93[Q/L/E]Add order by to test
ef472020-01-030.92[Q/L/E]Combine hashes using multiplication and XOR instead of only XOR
7bf92020-01-030.94[Q/L/E]JoinHashTable: switch to using Select comparisons for probes
f27e2020-01-030.92[Q/L/E]Eliminate duplicate expressions in join order optimizer and fix cost of plans
d58a2020-01-020.92[Q/L/E]Simplify code for SINGLE join
76c22020-01-020.93[Q/L/E]Revert "Use Select comparisons instead of regular comparisons in join ht" This reverts commit eccbf85d36f558ff65c37f865dad0d24325f389d.
eccb2020-01-020.92[Q/L/E]Use Select comparisons instead of regular comparisons in join ht
33642020-01-020.92[Q/L/E]Merge branch 'master' into scan
6d1a2020-01-020.93[Q/L/E]Make join hashtable allocate blocks from the buffer manager instead of allocating memory directly
4b0b2020-01-020.92[Q/L/E]Remove unused function and variable
70c22020-01-020.93[Q/L/E]Modify hash to use a single multiplication instead of using the murmurhash
ebee2020-01-020.92[Q/L/E]Correctly handle empty HT for LEFT and SINGLE join
9a892020-01-020.92[Q/L/E]No longer resize join HT, instead only construct the join HT after all chunks have been computed
f75c2020-01-020.93[Q/L/E]Add ORDER BY to left join test
10042020-01-010.93[Q/L/E]Add cache after hash table probe that re-probes and caches if there are too few matches to avoid outputting small chunks This time to the correct branch
ffe92020-01-010.93[Q/L/E]Revert "Add cache after hash table probe that re-probes and caches if there are too few matches to avoid outputting small chunks" This reverts commit d9a40fa733c9b7365653cb5d4c4cfd7fd382e14c.
d9a42020-01-010.93[Q/L/E]Add cache after hash table probe that re-probes and caches if there are too few matches to avoid outputting small chunks
ce842020-01-010.93[Q/L/E]Fix path to js files in
fe312019-12-310.93[Q/L/E]Merge pull request #382 from cwida/between Add BETWEEN operator and rework OR/(NOT) IN operators
f6eb2019-12-310.93[Q/L/E]Fix copy ellision warning
caa32019-12-310.93[Q/L/E]Also rewrite NOT IN, to either a MARK join (in case of many elements) or (X <> 1 AND X <> 2 AND ...)
bd332019-12-310.93[Q/L/E]Convert IN clause with few children into OR expression (i.e. X=1 OR X=2 OR X=3...)
35452019-12-310.93[Q/L/E]Rework OR statement in FILTER to use selection vector to exclude tuples that have already qualified
d9fc2019-12-300.94[Q/L/E]Replace GetNumericValue with GetValue
03cb2019-12-300.93[Q/L/E]Clean up unused unary loops
e4742019-12-300.93[Q/L/E]Remove parsed BetweenExpression and turn BETWEEN in parse tree into two separate comparison expressions again - they will later be turned into a BoundBetweenExpression again
e5352019-12-300.94[Q/L/E]In FilterCombiner, generate a BETWEEN expression when there is both a lower and upper bound for a variable
7d302019-12-300.93[Q/L/E]More BETWEEN tests
713d2019-12-300.93[Q/L/E]Add BETWEEN and BOUND BETWEEN expressions
af3c2019-12-300.93[Q/L/E]Add ternary loops and cleanup binary loops, plus add more tests for BETWEEN
80912019-12-290.93[Q/L/E]Add --query option to benchmark_runner
5df62019-12-280.92[Q/L/E]Merge pull request #380 from cwida/benchmarkdatareuse Efficient CSV Reader and Benchmark Data Re-Use
0c4d2019-12-280.93[Q/L/E]One more CSV edge case
0e972019-12-280.92[Q/L/E]Add more tests for several CSV reader edge cases
74612019-12-280.94[Q/L/E]Re-write complex CSV reader (that supports multi-byte delimiters/escapes/quotes) to be more efficient
4bfe2019-12-282.44[Q/L/E]Simplify CSV reader somewhat to avoid duplicate code
a0c42019-12-282.44[Q/L/E]Add benchmark for reading a CSV with multi-byte delimiters