https://github.com/rishiloyola/d-left
data structure to store data, rapidly, memory-efficiently and with less collision
https://github.com/rishiloyola/d-left
bloom-filter d-left hashing kernel
Last synced: 16 days ago
JSON representation
data structure to store data, rapidly, memory-efficiently and with less collision
- Host: GitHub
- URL: https://github.com/rishiloyola/d-left
- Owner: rishiloyola
- License: mit
- Created: 2016-02-12T12:19:16.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2017-04-09T09:52:36.000Z (about 8 years ago)
- Last Synced: 2025-05-08T00:53:23.316Z (16 days ago)
- Topics: bloom-filter, d-left, hashing, kernel
- Language: C
- Homepage:
- Size: 11.7 KB
- Stars: 5
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# d-left
A d-left is a data structure designed to store data, rapidly, memory-efficiently and with very less collision probability compare to Bloom Filter. Check out this [experiment](https://www.eecs.harvard.edu/~michaelm/postscripts/aller2006.pdf) to increase the capacity and to reduce the false positive of Bloom Filter.
### Advantages :
* Separation of data and use of fingerprints allows checking of several locations per memory access.
* Multiple hash functions significatly reduces collision probability.
* Number of memory accesses rarely exceeds d+1.
### Run
```
gcc -c murmur3.c hash.c main.c
gcc -o test main.o murmur3.o
./test
```
### Dependency
[Murmur3 hashing](https://github.com/PeterScott/murmur3)
### Refereces
1. [Algorithms - ESA 2006](https://books.google.co.in/books?id=eKOoCAAAQBAJ&pg=PA688&lpg=PA688&dq=d-left+deletion+hashing&source=bl&ots=_t4AgfeEqs&sig=6pc27jDRL5SybempvGOaUr62jlE&hl=en&sa=X&redir_esc=y#v=onepage&q=d-left%20hashing&f=false)
2. [Slide](http://www.arl.wustl.edu/~jst/cse/577/lec/exactMatch.pdf)
### Usage
```c
int main(int argc, char const *argv[])
{
struct Table T1;
init(&T1);
bool output;
for (int i = 0; i < 10; i++) {
output = inserting(i, &T1);
}
output = delete(5, &T1);
output = delete(1, &T1);
output = lookup(5, &T1);
return 0;
}
```
### Config
Change the below parameters to generate the hash table of suitable size.
```c
const int BUCKET_HEIGHT = 2;
const int SUBTABLE_SIZE = 5;
const int TABLE_SIZE = 13;
const uint32_t SEED = 42;
```
##### Note:
* Currently it only supports **uint32_t** data type.