1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
|
<!-- README.md is generated from README.Rmd. Please edit that file -->
# sparseMatrixStats <a href='https://github.com/const-ae/sparseMatrixStats'><img src='man/figures/logo.png' align="right" height="209" /></a>
<!-- badges: start -->
[](https://codecov.io/gh/const-ae/sparseMatrixStats)
<!-- badges: end -->
The goal of `sparseMatrixStats` is to make the API of
[matrixStats](https://github.com/HenrikBengtsson/matrixStats) available
for sparse matrices.
## Installation
You can install the release version of
*[sparseMatrixStats](https://bioconductor.org/packages/sparseMatrixStats)*
from BioConductor:
``` r
if (!requireNamespace("BiocManager", quietly = TRUE))
install.packages("BiocManager")
BiocManager::install("sparseMatrixStats")
```
Alternatively, you can get the development version of the package from
[GitHub](https://github.com/const-ae/sparseMatrixStats) with:
``` r
# install.packages("devtools")
devtools::install_github("const-ae/sparseMatrixStats")
```
## Example
``` r
library(sparseMatrixStats)
```
``` r
mat <- matrix(0, nrow=10, ncol=6)
mat[sample(seq_len(60), 4)] <- 1:4
# Convert dense matrix to sparse matrix
sparse_mat <- as(mat, "dgCMatrix")
sparse_mat
#> 10 x 6 sparse Matrix of class "dgCMatrix"
#>
#> [1,] 4 . . . . .
#> [2,] . . . . . .
#> [3,] . . . . . .
#> [4,] 2 . . . . .
#> [5,] . . . . . .
#> [6,] . . . . . .
#> [7,] . . . . . 1
#> [8,] . . . . . .
#> [9,] . . . 3 . .
#> [10,] . . . . . .
```
The package provides an interface to quickly do common operations on the
rows or columns. For example calculate the variance:
``` r
apply(mat, 2, var)
#> [1] 1.822222 0.000000 0.000000 0.900000 0.000000 0.100000
matrixStats::colVars(mat)
#> [1] 1.822222 0.000000 0.000000 0.900000 0.000000 0.100000
sparseMatrixStats::colVars(sparse_mat)
#> [1] 1.822222 0.000000 0.000000 0.900000 0.000000 0.100000
```
On this small example data, all methods are basically equally fast, but
if we have a much larger dataset, the optimizations for the sparse data
start to show.
I generate a dataset with 10,000 rows and 50 columns that is 99% empty
``` r
big_mat <- matrix(0, nrow=1e4, ncol=50)
big_mat[sample(seq_len(1e4 * 50), 5000)] <- rnorm(5000)
# Convert dense matrix to sparse matrix
big_sparse_mat <- as(big_mat, "dgCMatrix")
```
I use the `bench` package to benchmark the performance difference:
``` r
bench::mark(
sparseMatrixStats=sparseMatrixStats::colVars(big_sparse_mat),
matrixStats=matrixStats::colVars(big_mat),
apply=apply(big_mat, 2, var)
)
#> # A tibble: 3 x 6
#> expression min median `itr/sec` mem_alloc `gc/sec`
#> <bch:expr> <bch:tm> <bch:tm> <dbl> <bch:byt> <dbl>
#> 1 sparseMatrixStats 36.15µs 40.09µs 24419. 2.93KB 14.7
#> 2 matrixStats 1.42ms 1.45ms 677. 156.8KB 2.03
#> 3 apply 8.89ms 10.56ms 94.6 9.54MB 53.0
```
As you can see `sparseMatrixStats` is ca. 35 times fast than
`matrixStats`, which in turn is 7 times faster than the `apply()`
version.
# API
The package now supports all functions from the `matrixStats` API for
column sparse matrices (`dgCMatrix`). And thanks to the
[`MatrixGenerics`](https://bioconductor.org/packages/MatrixGenerics/) it
can be easily integrated along-side
[`matrixStats`](https://cran.r-project.org/package=matrixStats) and
[`DelayedMatrixStats`](https://bioconductor.org/packages/DelayedMatrixStats/).
Note that the `rowXXX()` functions are called by transposing the input
and calling the corresponding `colXXX()` function. Special optimized
implementations are available for `rowSums2()`, `rowMeans2()`, and
`rowVars()`.
| Method | matrixStats | sparseMatrixStats | Notes |
| :------------------- | :---------- | :---------------- | :--------------------------------------------------------------------------------------- |
| colAlls() | ✔ | ✔ | |
| colAnyMissings() | ✔ | ❌ | Not implemented because it is deprecated in favor of `colAnyNAs()` |
| colAnyNAs() | ✔ | ✔ | |
| colAnys() | ✔ | ✔ | |
| colAvgsPerRowSet() | ✔ | ✔ | |
| colCollapse() | ✔ | ✔ | |
| colCounts() | ✔ | ✔ | |
| colCummaxs() | ✔ | ✔ | |
| colCummins() | ✔ | ✔ | |
| colCumprods() | ✔ | ✔ | |
| colCumsums() | ✔ | ✔ | |
| colDiffs() | ✔ | ✔ | |
| colIQRDiffs() | ✔ | ✔ | |
| colIQRs() | ✔ | ✔ | |
| colLogSumExps() | ✔ | ✔ | |
| colMadDiffs() | ✔ | ✔ | |
| colMads() | ✔ | ✔ | |
| colMaxs() | ✔ | ✔ | |
| colMeans2() | ✔ | ✔ | |
| colMedians() | ✔ | ✔ | |
| colMins() | ✔ | ✔ | |
| colOrderStats() | ✔ | ✔ | |
| colProds() | ✔ | ✔ | |
| colQuantiles() | ✔ | ✔ | |
| colRanges() | ✔ | ✔ | |
| colRanks() | ✔ | ✔ | |
| colSdDiffs() | ✔ | ✔ | |
| colSds() | ✔ | ✔ | |
| colsum() | ✔ | ❌ | Base R function |
| colSums2() | ✔ | ✔ | |
| colTabulates() | ✔ | ✔ | |
| colVarDiffs() | ✔ | ✔ | |
| colVars() | ✔ | ✔ | |
| colWeightedMads() | ✔ | ✔ | Sparse version behaves slightly differently, because it always uses `interpolate=FALSE`. |
| colWeightedMeans() | ✔ | ✔ | |
| colWeightedMedians() | ✔ | ✔ | Only equivalent if `interpolate=FALSE` |
| colWeightedSds() | ✔ | ✔ | |
| colWeightedVars() | ✔ | ✔ | |
| rowAlls() | ✔ | ✔ | |
| rowAnyMissings() | ✔ | ❌ | Not implemented because it is deprecated in favor of `rowAnyNAs()` |
| rowAnyNAs() | ✔ | ✔ | |
| rowAnys() | ✔ | ✔ | |
| rowAvgsPerColSet() | ✔ | ✔ | |
| rowCollapse() | ✔ | ✔ | |
| rowCounts() | ✔ | ✔ | |
| rowCummaxs() | ✔ | ✔ | |
| rowCummins() | ✔ | ✔ | |
| rowCumprods() | ✔ | ✔ | |
| rowCumsums() | ✔ | ✔ | |
| rowDiffs() | ✔ | ✔ | |
| rowIQRDiffs() | ✔ | ✔ | |
| rowIQRs() | ✔ | ✔ | |
| rowLogSumExps() | ✔ | ✔ | |
| rowMadDiffs() | ✔ | ✔ | |
| rowMads() | ✔ | ✔ | |
| rowMaxs() | ✔ | ✔ | |
| rowMeans2() | ✔ | ✔ | |
| rowMedians() | ✔ | ✔ | |
| rowMins() | ✔ | ✔ | |
| rowOrderStats() | ✔ | ✔ | |
| rowProds() | ✔ | ✔ | |
| rowQuantiles() | ✔ | ✔ | |
| rowRanges() | ✔ | ✔ | |
| rowRanks() | ✔ | ✔ | |
| rowSdDiffs() | ✔ | ✔ | |
| rowSds() | ✔ | ✔ | |
| rowsum() | ✔ | ❌ | Base R function |
| rowSums2() | ✔ | ✔ | |
| rowTabulates() | ✔ | ✔ | |
| rowVarDiffs() | ✔ | ✔ | |
| rowVars() | ✔ | ✔ | |
| rowWeightedMads() | ✔ | ✔ | Sparse version behaves slightly differently, because it always uses `interpolate=FALSE`. |
| rowWeightedMeans() | ✔ | ✔ | |
| rowWeightedMedians() | ✔ | ✔ | Only equivalent if `interpolate=FALSE` |
| rowWeightedSds() | ✔ | ✔ | |
| rowWeightedVars() | ✔ | ✔ | |
|
