scriptoriumadmin/featurebase
1
0
Fork 0
mirror of https://github.com/featurebasedb/featurebase.git synced 2026-09-06 08:35:55 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
1677 commits 45 branches 174 tags 85 MiB
Commit graph

83 commits

Author SHA1 Message Date
Matt Jaffee
8ec8fc556c intersectRunRun to use runAppendInterval 
this avoid situations where two runs which are adjacent don't get collapsed to a
single run
 2017-06-19 10:13:02 -05:00
Matt Jaffee
97346baa94 unionArrayRun implementation 2017-06-19 10:13:02 -05:00
Matt Jaffee
c5a0f23257 clean up unionRunRun 2017-06-19 10:13:02 -05:00
Matt Jaffee
4848a99aca unionRunRun working - messy 2017-06-19 10:13:02 -05:00
Matt Jaffee
24c03f07f6 add note to improve intersectRunRun func 2017-06-19 10:13:02 -05:00
Matt Jaffee
cf445f8012 stub out the rest of the rle methods 2017-06-19 10:13:02 -05:00
Matt Jaffee
9781a4c78b fix tests, prep for union rle implementations 2017-06-19 10:13:02 -05:00
Matt Jaffee
250e8f6f5b implement intersectBitmapRun 
wowza
 2017-06-19 10:13:02 -05:00
Matt Jaffee
8954a6ae8c fix bitmapCountRange and test 
had been thinking that index 0 was the most significant bit, but based on the
bitmapAdd function, it must be the least significant bit
 2017-06-19 10:13:02 -05:00
Matt Jaffee
3e8030aee4 implement intersectRunRun 2017-06-19 10:10:57 -05:00
Matt Jaffee
0c3f61a426 implement intersectArrayRun 2017-06-19 10:10:57 -05:00
Matt Jaffee
20cec1363b implement intersectionCounts for run containers 2017-06-19 10:10:57 -05:00
Matt Jaffee
014005b5d5 when choosing between container types, keep bitmap as default 2017-06-19 10:06:11 -05:00
Matt Jaffee
95b55ee7db use uint32 instead of uint16 for rle containers 
stays consistent with the array type containers. array and rle are using uint32
in case we want to change the size of roaring containers as an optimization.
This has been done in the past
a8feb985e1

The downside is that more storage is used, so we may want to convert to uint16
until we decide to play with that optimization again, but for now I'll keep the
rle implementation in sync with the array implementation.
 2017-06-19 10:06:11 -05:00
Matt Jaffee
98c8fd1ded rename conversion funcs to prepare for rle 
also support rle in clone()
 2017-06-19 10:06:11 -05:00
Matt Jaffee
e80812cb0a implement runMax 2017-06-19 10:06:11 -05:00
Matt Jaffee
72b6eb45e5 add runRemove 2017-06-19 10:06:11 -05:00
Matt Jaffee
c2203d19c1 wip on adding RLE to roaring, add, countRange, contains 2017-06-19 10:03:55 -05:00
Ben Johnson
ca29f7692b
Fix failing tests when IPv6 is disabled. 
Turning off IPv6 causes the tests to fail because the Go
`net.Listen()` function obtains a `[::]` address when using
a `localhost:0` bind address. Looking through the Go source
code, I don't see a good way to fix that.

This change adds a test flag to allow users to specify the
network name when running the test:

```sh
go test github.com/pilosa/pilosa/server -network tcp4
```
 2017-06-01 14:43:59 -06:00
Matthew Jaffee
fa4234b04c Merge pull request #590 from jaffee/set-n-at-end 
set n based on array length instead of incrementing repeatedly
 2017-06-01 11:03:35 -05:00
Todd Gruben
542d748baf corrected tests to no longer use String method 2017-06-01 10:17:49 -05:00
Todd Gruben
ba2b8cd5b1 removed String method upon discovery of %v 2017-06-01 10:03:23 -05:00
Travis
879e02de81
fix misspelling and run gofmt 2017-05-31 16:22:43 -05:00
Todd Gruben
f584400330 attempt to conquer weird 2017-05-31 16:10:48 -05:00
Todd Gruben
e12fdb69a1 better String() 2017-05-31 15:04:33 -05:00
Todd Gruben
7382428f38 clensed String method;added documentation for tests 2017-05-31 15:01:48 -05:00
Todd Gruben
649e58253a corrected flip prefix;changed flip to inclusive 2017-05-31 13:58:54 -05:00
Todd Gruben
a6975192d6 Merge branch 'flip' of github.com:tgruben/pilosa into flip 2017-05-31 10:55:31 -05:00
Todd Gruben
aa13bbb6ce cleaned up comments 2017-05-31 10:40:52 -05:00
tgruben
a58206bb4d Merge branch 'master' into flip 2017-05-31 08:43:50 -05:00
tgruben
8c5c48bba9 Merge pull request #571 from tgruben/xor 
added xor support to roaring
 2017-05-31 08:34:58 -05:00
Matt Jaffee
56ba127210 set n based on array length instead of incrementing repeatedly 2017-05-30 16:09:02 -05:00
Todd Gruben
1c468fb38e bitmap container xor empty test 2017-05-30 15:07:10 -05:00
Todd Gruben
b5bd8433cc added empty and equivlence testing 2017-05-30 13:27:11 -05:00
Todd Gruben
7090ee4d7a @jaffee cleanup 2017-05-30 12:39:23 -05:00
Todd Gruben
498045f25a added copy optimization ; type check; and code cleanup 2017-05-30 11:18:05 -05:00
Todd Gruben
aa713e3832 add support for bit flip (negate) in roaring 2017-05-25 10:07:29 -05:00
Matt Jaffee
da961be614 rewrite intersectCountArrayBitmap for perf test 2017-05-24 12:38:50 -05:00
Todd Gruben
2182534cc7 added xor support to roaring 2017-05-22 16:51:53 -05:00
Matt Jaffee
e90e1455db remove branch and always count last container 2017-05-22 10:38:40 -05:00
Matt Jaffee
6b3b459659 add license to roaring_internal_test 2017-05-22 09:46:52 -05:00
Matt Jaffee
a5918d01d9 fix offset check to make sense 2017-05-19 17:09:22 -05:00
Matt Jaffee
c0ddbe0e3f fix bitmapCountRange and test 
had been thinking that index 0 was the most significant bit, but based on the
bitmapAdd function, it must be the least significant bit
 2017-05-19 15:18:07 -05:00
Matt Jaffee
fb4651cdb8 fix 3 separate bugs in bitmapCountRange 
in order of the diff:

1. When the start and end of the range fall in the same word, special handling
is needed to "mask" off the beginning and end of the word simultaneously to
avoid counting bits at the beginning or end of the word that aren't in the
range.

2. `i++` is needed at the end of the first partial word to avoid counting this
word in the next block.

3. the shift amount for right shifts is 64 - (end % 64) rather than just end
% 64. If end is (e.g.) 68, then 68 - 64 is 4 and we are only interested in the
first 4 bits of the word, so we must right shift by 60 bits, not 4 bits.
 2017-05-18 12:04:37 -05:00
Todd Gruben
fbae0f4d15 asm build options for non amd64 2017-05-15 11:31:41 -05:00
Damian Gryski
62d6edd1ca roaring: fix vet issues with the assembly code 
assembly_amd64.s:3: [amd64] hasAsm: wrong argument size 0; expected $...-1
assembly_amd64.s:11: [amd64] POPCNTQ: wrong argument size 8; expected $...-16
assembly_amd64.s:12: [amd64] POPCNTQ: unknown variable x; offset 0 is memory+0(FP)
assembly_amd64.s:17: [amd64] BSFQ: wrong argument size 8; expected $...-16
assembly_amd64.s:18: [amd64] BSFQ: unknown variable x; offset 0 is memory+0(FP)
assembly_amd64.s:28: [amd64] popcntSliceAsm: invalid offset s+8(FP); expected s+0(FP), s_base+0(FP), s_len+8(FP), or s_cap+16(FP)
assembly_amd64.s:43: [amd64] popcntMaskSliceAsm: invalid offset s+8(FP); expected s+0(FP), s_base+0(FP), s_len+8(FP), or s_cap+16(FP)
assembly_amd64.s:63: [amd64] popcntAndSliceAsm: invalid offset s+8(FP); expected s+0(FP), s_base+0(FP), s_len+8(FP), or s_cap+16(FP)
assembly_amd64.s:82: [amd64] popcntOrSliceAsm: invalid offset s+8(FP); expected s+0(FP), s_base+0(FP), s_len+8(FP), or s_cap+16(FP)
assembly_amd64.s:101: [amd64] popcntXorSliceAsm: invalid offset s+8(FP); expected s+0(FP), s_base+0(FP), s_len+8(FP), or s_cap+16(FP)
 2017-05-04 11:01:10 +02:00
Travis
ea4bf0a1d3
Fix roaring test: TestBitmap_Quick_Array1 
When the random number generator was coming up 0,
then `a = []uint64{}`. This caused the bitmap `bm` to be empty.
In this case, `bm.Slice()` is a nil slice, while the expected
slice is an empty slice (i.e. they are not considered equal):
got: ([]uint64) <nil>
exp: ([]uint64) {}
 2017-04-29 18:04:59 -05:00
Cody Soyland
3996b56b88 Apply Apache License 2.0 2017-04-28 14:22:03 -05:00
Travis
c3ed12c20e Refactor fragment.bitmap() so that it leverages bitmapCache and so that 
it's no longer reponsible for updating the count cache.
This commit also helps SetBit/ClearBit performance by allowing them
to work against data from `bitmapCache` instead of loading bitmaps
from fragment.storage every time.
 2017-03-06 11:06:22 -06:00
Todd Gruben
43ede4d9c8 WIP increase caching and topn adjustments 2017-02-21 09:28:01 -06:00