# kwgoodman/bottleneck

Fast NumPy array functions written in C

Language: C

Keywords: c, c-extension, fast, numpy, python

## Bottleneck

Bottleneck is a collection of fast NumPy array functions written in C.

Let's give it a try. Create a NumPy array:

```>>> import numpy as np
>>> a = np.array([1, 2, np.nan, 4, 5])
```

Find the nanmean:

```>>> import bottleneck as bn
>>> bn.nanmean(a)
3.0
```

Moving window mean:

```>>> bn.move_mean(a, window=2, min_count=1)
array([ 1. ,  1.5,  2. ,  4. ,  4.5])
```

### Benchmark

Bottleneck comes with a benchmark suite:

```>>> bn.bench()
Bottleneck performance benchmark
Bottleneck 1.3.0.dev0; Numpy 1.12.1
Speed is NumPy time divided by Bottleneck time
NaN means approx one-fifth NaNs; float64 used

no NaN     no NaN      NaN       no NaN      NaN
(100,)  (1000,1000)(1000,1000)(1000,1000)(1000,1000)
axis=0     axis=0     axis=0     axis=1     axis=1
nansum         67.3        0.3        0.7        2.5        2.4
nanmean       194.8        1.9        2.1        3.4        3.1
nanstd        241.5        1.6        2.1        2.7        2.6
nanvar        229.7        1.7        2.1        2.7        2.5
nanmin         34.1        0.7        1.1        0.8        2.6
nanmax         45.6        0.7        2.7        1.0        3.7
median        111.0        1.3        5.6        1.0        4.8
nanmedian     108.8        5.9        6.7        5.6        6.7
ss             16.3        1.1        1.2        1.6        1.6
nanargmin      89.2        2.9        5.1        2.2        5.6
nanargmax     107.4        3.0        5.4        2.2        5.8
anynan         19.4        0.3       35.0        0.5       29.9
allnan         39.9      146.6      128.3      115.8       75.6
rankdata       55.0        2.6        2.3        2.9        2.8
nanrankdata    59.8        2.8        2.2        3.2        2.5
partition       4.4        1.2        1.6        1.0        1.4
argpartition    3.5        1.1        1.4        1.1        1.6
replace        17.7        1.4        1.4        1.3        1.4
push         3440.0        7.8        9.5       20.0       15.5
move_sum     4743.0       75.7      156.1      195.4      211.1
move_mean    8760.9      116.2      167.4      252.1      258.8
move_std     8979.9       96.1      196.3      144.0      256.3
move_var    11216.8      127.3      243.9      225.9      321.4
move_min     2245.3       20.6       36.7       23.2       42.1
move_max     2223.7       20.5       37.2       24.1       42.4
move_argmin  3664.0       48.2       73.3       40.2       83.9
move_argmax  3916.9       42.0       75.4       41.5       81.2
move_median  2023.3      166.8      173.7      153.8      154.3
move_rank    1208.5        1.9        1.9        2.5        2.8
```

You can also run a detailed benchmark for a single function using, for example, the command:

```>>> bn.bench_detailed("move_median", fraction_nan=0.3)
```

Only arrays with data type (dtype) int32, int64, float32, and float64 are accelerated. All other dtypes result in calls to slower, unaccelerated functions. In the rare case of a byte-swapped input array (e.g. a big-endian array on a little-endian operating system) the function will not be accelerated regardless of dtype.

### Install

Requirements:

 Bottleneck Python 2.7, 3.5, 3.6; NumPy 1.12.1 Compile gcc, clang, MinGW or MSVC Unit tests nose

To install Bottleneck on GNU/Linux, Mac OS X, et al.:

```\$ sudo python setup.py install
```

To install bottleneck on Windows, first install MinGW and add it to your system path. Then install Bottleneck with the commands:

```python setup.py install --compiler=mingw32
```

Alternatively, you can use the Windows binaries created by Christoph Gohlke: http://www.lfd.uci.edu/~gohlke/pythonlibs/#bottleneck

### Unit tests

After you have installed Bottleneck, run the suite of unit tests:

```>>> import bottleneck as bn
>>> bn.test()
<snip>
Ran 169 tests in 57.205s
OK
<nose.result.TextTestResult run=169 errors=0 failures=0>
```

#### Project Statistics

 Sourcerank 10 Repository Size 12.5 MB Stars 256 Forks 37 Watchers 16 Open issues 10 Dependencies 2 Contributors 13 Tags 21 Created Nov 27, 2010 Last updated Apr 6, 2018 Last pushed Jan 9, 2018

#### Packages Referencing this Repo

##### Bottleneck
Fast NumPy array functions written in C
Latest release 1.2.1 - Updated - 256 stars

#### Recent Tags See all

 v1.2.1 May 15, 2017 v1.2.0 October 20, 2016 v1.1.0 June 22, 2016 v1.0.0 February 06, 2015 v0.8.0 January 21, 2014 v0.7.0 September 10, 2013 v0.6.0 June 04, 2012 v0.6rc1 May 23, 2012 v0.5.0 June 13, 2011 v0.5.0rc3 June 10, 2011 v0.5.0rc2 June 10, 2011 v0.5.0rc1 June 07, 2011 v0.5.0beta2 June 05, 2011 v0.5.0beta June 04, 2011 v0.4.3 March 17, 2011

#### Interesting Forks See all

##### dougalsutherland/bottleneck
Fast NumPy array functions written in Cython
Python - Other - Last pushed - 3 stars
##### stroxler/bottleneck
Fast NumPy array functions written in Cython
C - Other - Last pushed - 2 stars
##### fhal/bottleneck
Fast NumPy array functions written in Cython
Python - Other - Last pushed - 2 stars
##### WeatherGod/bottleneck
Fast NumPy array functions written in Cython
Python - Other - Last pushed - 2 stars
##### ml31415/bottleneck
Fast NumPy array functions written in Cython
Python - Other - Last pushed - 1 stars