Fast (Grouped, Weighted) Sum for Matrix-Like Objects
fsum.Rdfsum is a generic function that computes the (column-wise) sum of all values in x, (optionally) grouped by g and/or weighted by w (e.g. to calculate survey totals). The TRA argument can further be used to transform x using its (grouped, weighted) sum.
Usage
fsum(x, ...)
# Default S3 method
fsum(x, g = NULL, w = NULL, TRA = NULL, na.rm = .op[["na.rm"]],
use.g.names = TRUE, fill = FALSE, nthreads = .op[["nthreads"]], ...)
# S3 method for class 'matrix'
fsum(x, g = NULL, w = NULL, TRA = NULL, na.rm = .op[["na.rm"]],
use.g.names = TRUE, drop = TRUE, fill = FALSE, nthreads = .op[["nthreads"]], ...)
# S3 method for class 'data.frame'
fsum(x, g = NULL, w = NULL, TRA = NULL, na.rm = .op[["na.rm"]],
use.g.names = TRUE, drop = TRUE, fill = FALSE, nthreads = .op[["nthreads"]], ...)
# S3 method for class 'grouped_df'
fsum(x, w = NULL, TRA = NULL, na.rm = .op[["na.rm"]],
use.g.names = FALSE, keep.group_vars = TRUE, keep.w = TRUE, stub = .op[["stub"]],
fill = FALSE, nthreads = .op[["nthreads"]], ...)Arguments
- x
a numeric vector, matrix, data frame or grouped data frame (class 'grouped_df').
- g
a factor,
GRPobject, atomic vector (internally converted to factor) or a list of vectors / factors (internally converted to aGRPobject) used to groupx.- w
a numeric vector of (non-negative) weights, may contain missing values.
- TRA
an integer or quoted operator indicating the transformation to perform: 0 - "na" | 1 - "fill" | 2 - "replace" | 3 - "-" | 4 - "-+" | 5 - "/" | 6 - "%" | 7 - "+" | 8 - "*" | 9 - "%%" | 10 - "-%%". See
TRA.- na.rm
logical. Skip missing values in
x. Defaults toTRUEand implemented at very little computational cost. Ifna.rm = FALSEaNAis returned when encountered.- use.g.names
logical. Make group-names and add to the result as names (default method) or row-names (matrix and data frame methods). No row-names are generated for data.table's.
- fill
logical. Initialize result with
0instead ofNAwhenna.rm = TRUEe.g.fsum(NA, fill = TRUE)returns0instead ofNA.- nthreads
integer. The number of threads to utilize. See Details.
- drop
matrix and data.frame method: Logical.
TRUEdrops dimensions and returns an atomic vector ifg = NULLandTRA = NULL.- keep.group_vars
grouped_df method: Logical.
FALSEremoves grouping variables after computation.- keep.w
grouped_df method: Logical. Retain summed weighting variable after computation (if contained in
grouped_df).- stub
character. If
keep.w = TRUEandstub = TRUE(default), the summed weights column is prefixed by"sum.". Users can specify a different prefix through this argument, or set it toFALSEto avoid prefixing.- ...
arguments to be passed to or from other methods. If
TRAis used, passingset = TRUEwill transform data by reference and return the result invisibly.
Details
The weighted sum (e.g. survey total) is computed as sum(x * w), but in one pass and about twice as efficient. If na.rm = TRUE, missing values will be removed from both x and w i.e. utilizing only x[complete.cases(x,w)] and w[complete.cases(x,w)].
This all seamlessly generalizes to grouped computations, which are performed in a single pass (without splitting the data) and are therefore extremely fast. See Benchmark and Examples below.
When applied to data frames with groups or drop = FALSE, fsum preserves all column attributes. The attributes of the data frame itself are also preserved.
Since v1.6.0 fsum explicitly supports integers. Integers are summed using the long long type in C which is bounded at +-9,223,372,036,854,775,807 (so ~4.3 billion times greater than the minimum/maximum R integer bounded at +-2,147,483,647). If the value of the sum is outside +-2,147,483,647, a double containing the result is returned, otherwise an integer is returned. With groups, an integer results vector is initialized, and an integer overflow error is provided if the sum in any group is outside +-2,147,483,647. Data needs to be coerced to double beforehand in such cases.
Multithreading, added in v1.8.0, applies at the column-level unless g = NULL and nthreads > NCOL(x). Parallelism over groups is not available because sums are computed simultaneously within each group. nthreads = 1L uses a serial version of the code, not parallel code running on one thread. This serial code is always used with less than 100,000 obs (length(x) < 100000 for vectors and matrices), because parallel execution itself has some overhead.
Value
The (w weighted) sum of x, grouped by g, or (if TRA is used) x transformed by its (grouped, weighted) sum.
Examples
## default vector method
mpg <- mtcars$mpg
fsum(mpg) # Simple sum
#> [1] 642.9
fsum(mpg, w = mtcars$hp) # Weighted sum (total): Weighted by hp
#> [1] 84362.7
fsum(mpg, TRA = "%") # Simple transformation: obtain percentages of mpg
#> [1] 3.266449 3.266449 3.546430 3.328667 2.908695 2.815368 2.224296 3.795303
#> [9] 3.546430 2.986468 2.768704 2.550941 2.690932 2.364287 1.617670 1.617670
#> [17] 2.286514 5.039664 4.728574 5.272982 3.344221 2.410950 2.364287 2.068751
#> [25] 2.986468 4.246384 4.044175 4.728574 2.457614 3.064240 2.333178 3.328667
fsum(mpg, mtcars$cyl) # Grouped sum
#> 4 6 8
#> 293.3 138.2 211.4
fsum(mpg, mtcars$cyl, mtcars$hp) # Weighted grouped sum (total)
#> 4 6 8
#> 23743.0 16873.0 43746.7
fsum(mpg, mtcars[c(2,8:9)]) # More groups..
#> 4.0.1 4.1.0 4.1.1 6.0.1 6.1.0 8.0.0 8.0.1
#> 26.0 68.7 198.6 61.7 76.5 180.6 30.8
g <- GRP(mtcars, ~ cyl + vs + am) # Precomputing groups gives more speed !
fsum(mpg, g)
#> 4.0.1 4.1.0 4.1.1 6.0.1 6.1.0 8.0.0 8.0.1
#> 26.0 68.7 198.6 61.7 76.5 180.6 30.8
fmean(mpg, g) == fsum(mpg, g) / fnobs(mpg, g)
#> 4.0.1 4.1.0 4.1.1 6.0.1 6.1.0 8.0.0 8.0.1
#> TRUE TRUE TRUE TRUE TRUE TRUE TRUE
fsum(mpg, g, TRA = "%") # Percentages by group
#> [1] 34.035656 34.035656 11.480363 27.973856 10.354374 23.660131
#> [7] 7.918051 35.516739 33.187773 25.098039 23.267974 9.080842
#> [13] 9.579181 8.416390 5.758583 5.758583 8.139535 16.314199
#> [19] 15.307150 17.069486 31.295488 8.582503 8.416390 7.364341
#> [25] 10.631229 13.746224 100.000000 15.307150 51.298701 31.928687
#> [31] 48.701299 10.775428
## data.frame method
fsum(mtcars)
#> mpg cyl disp hp drat wt qsec vs
#> 642.900 198.000 7383.100 4694.000 115.090 102.952 571.160 14.000
#> am gear carb
#> 13.000 118.000 90.000
fsum(mtcars, TRA = "%")
#> mpg cyl disp hp drat wt
#> Mazda RX4 3.266449 3.030303 2.167111 2.343417 3.388652 2.544875
#> Mazda RX4 Wag 3.266449 3.030303 2.167111 2.343417 3.388652 2.792564
#> Datsun 710 3.546430 2.020202 1.462800 1.981253 3.345208 2.253477
#> Hornet 4 Drive 3.328667 3.030303 3.494467 2.343417 2.676166 3.122815
#> Hornet Sportabout 2.908695 4.040404 4.876001 3.728164 2.736988 3.341363
#> Valiant 2.815368 3.030303 3.047500 2.236898 2.398123 3.360789
#> qsec vs am gear carb
#> Mazda RX4 2.881854 0.000000 7.692308 3.389831 4.444444
#> Mazda RX4 Wag 2.979901 0.000000 7.692308 3.389831 4.444444
#> Datsun 710 3.258281 7.142857 7.692308 3.389831 1.111111
#> Hornet 4 Drive 3.403600 7.142857 0.000000 2.542373 1.111111
#> Hornet Sportabout 2.979901 0.000000 0.000000 2.542373 2.222222
#> Valiant 3.540164 7.142857 0.000000 2.542373 1.111111
#> [ reached 'max' / getOption("max.print") -- omitted 26 rows ]
fsum(mtcars, g)
#> mpg cyl disp hp drat wt qsec vs am gear carb
#> 4.0.1 26.0 4 120.3 91 4.43 2.140 16.70 0 1 5 2
#> 4.1.0 68.7 12 407.6 254 11.31 8.805 62.91 3 0 11 5
#> 4.1.1 198.6 28 628.6 564 29.04 14.198 130.90 7 7 29 10
#> 6.0.1 61.7 18 465.0 395 11.42 8.265 48.98 0 3 13 14
#> 6.1.0 76.5 24 818.2 461 13.68 13.555 76.86 4 0 14 10
#> 8.0.0 180.6 96 4291.4 2330 37.45 49.249 205.71 0 0 36 37
#> [ reached 'max' / getOption("max.print") -- omitted 1 rows ]
fsum(mtcars, g, TRA = "%")
#> mpg cyl disp hp drat wt
#> Mazda RX4 34.03566 33.333333 34.408602 27.84810 34.150613 31.699940
#> Mazda RX4 Wag 34.03566 33.333333 34.408602 27.84810 34.150613 34.785239
#> Datsun 710 11.48036 14.285714 17.181037 16.48936 13.257576 16.340330
#> Hornet 4 Drive 27.97386 25.000000 31.532633 23.86117 22.514620 23.718185
#> Hornet Sportabout 10.35437 8.333333 8.388871 7.51073 8.411215 6.984913
#> Valiant 23.66013 25.000000 27.499389 22.77657 20.175439 25.525636
#> qsec vs am gear carb
#> Mazda RX4 33.605553 NaN 33.33333 30.769231 28.571429
#> Mazda RX4 Wag 34.748877 NaN 33.33333 30.769231 28.571429
#> Datsun 710 14.216960 14.28571 14.28571 13.793103 10.000000
#> Hornet 4 Drive 25.292740 25.00000 NaN 21.428571 10.000000
#> Hornet Sportabout 8.273783 NaN NaN 8.333333 5.405405
#> Valiant 26.307572 25.00000 NaN 21.428571 10.000000
#> [ reached 'max' / getOption("max.print") -- omitted 26 rows ]
## matrix method
m <- qM(mtcars)
fsum(m)
#> mpg cyl disp hp drat wt qsec vs
#> 642.900 198.000 7383.100 4694.000 115.090 102.952 571.160 14.000
#> am gear carb
#> 13.000 118.000 90.000
fsum(m, TRA = "%")
#> mpg cyl disp hp drat wt
#> Mazda RX4 3.266449 3.030303 2.1671114 2.343417 3.388652 2.544875
#> Mazda RX4 Wag 3.266449 3.030303 2.1671114 2.343417 3.388652 2.792564
#> Datsun 710 3.546430 2.020202 1.4628002 1.981253 3.345208 2.253477
#> Hornet 4 Drive 3.328667 3.030303 3.4944671 2.343417 2.676166 3.122815
#> Hornet Sportabout 2.908695 4.040404 4.8760006 3.728164 2.736988 3.341363
#> Valiant 2.815368 3.030303 3.0475004 2.236898 2.398123 3.360789
#> qsec vs am gear carb
#> Mazda RX4 2.881854 0.000000 7.692308 3.389831 4.444444
#> Mazda RX4 Wag 2.979901 0.000000 7.692308 3.389831 4.444444
#> Datsun 710 3.258281 7.142857 7.692308 3.389831 1.111111
#> Hornet 4 Drive 3.403600 7.142857 0.000000 2.542373 1.111111
#> Hornet Sportabout 2.979901 0.000000 0.000000 2.542373 2.222222
#> Valiant 3.540164 7.142857 0.000000 2.542373 1.111111
#> [ reached getOption("max.print") -- omitted 26 rows ]
fsum(m, g)
#> mpg cyl disp hp drat wt qsec vs am gear carb
#> 4.0.1 26.0 4 120.3 91 4.43 2.140 16.70 0 1 5 2
#> 4.1.0 68.7 12 407.6 254 11.31 8.805 62.91 3 0 11 5
#> 4.1.1 198.6 28 628.6 564 29.04 14.198 130.90 7 7 29 10
#> 6.0.1 61.7 18 465.0 395 11.42 8.265 48.98 0 3 13 14
#> 6.1.0 76.5 24 818.2 461 13.68 13.555 76.86 4 0 14 10
#> 8.0.0 180.6 96 4291.4 2330 37.45 49.249 205.71 0 0 36 37
#> [ reached getOption("max.print") -- omitted 1 row ]
fsum(m, g, TRA = "%")
#> mpg cyl disp hp drat
#> Mazda RX4 34.035656 33.333333 34.408602 27.848101 34.150613
#> Mazda RX4 Wag 34.035656 33.333333 34.408602 27.848101 34.150613
#> Datsun 710 11.480363 14.285714 17.181037 16.489362 13.257576
#> Hornet 4 Drive 27.973856 25.000000 31.532633 23.861171 22.514620
#> Hornet Sportabout 10.354374 8.333333 8.388871 7.510730 8.411215
#> Valiant 23.660131 25.000000 27.499389 22.776573 20.175439
#> wt qsec vs am gear
#> Mazda RX4 31.699940 33.605553 NaN 33.33333 30.769231
#> Mazda RX4 Wag 34.785239 34.748877 NaN 33.33333 30.769231
#> Datsun 710 16.340330 14.216960 14.28571 14.28571 13.793103
#> Hornet 4 Drive 23.718185 25.292740 25.00000 NaN 21.428571
#> Hornet Sportabout 6.984913 8.273783 NaN NaN 8.333333
#> Valiant 25.525636 26.307572 25.00000 NaN 21.428571
#> carb
#> Mazda RX4 28.571429
#> Mazda RX4 Wag 28.571429
#> Datsun 710 10.000000
#> Hornet 4 Drive 10.000000
#> Hornet Sportabout 5.405405
#> Valiant 10.000000
#> [ reached getOption("max.print") -- omitted 26 rows ]
## method for grouped data frames - created with dplyr::group_by or fgroup_by
mtcars |> fgroup_by(cyl,vs,am) |> fsum(hp) # Weighted grouped sum (total)
#> cyl vs am sum.hp mpg disp drat wt qsec gear carb
#> 1 4 0 1 91 2366.0 10947.3 403.13 194.740 1519.70 455 182
#> 2 4 1 0 254 5764.3 34121.1 960.08 736.135 5356.47 919 411
#> 3 4 1 1 564 15612.7 52945.4 2301.27 1165.914 10456.79 2369 838
#> 4 6 0 1 395 8067.5 60575.0 1491.50 1089.200 6395.30 1755 1930
#> 5 6 1 0 461 8805.5 93234.6 1592.92 1563.190 8837.10 1629 1199
#> 6 8 0 0 2330 34550.5 846042.0 7323.40 9689.735 39807.80 6990 7480
#> [ reached 'max' / getOption("max.print") -- omitted 1 rows ]
mtcars |> fgroup_by(cyl,vs,am) |> fsum(TRA = "%")
#> cyl vs am mpg disp hp drat wt
#> Mazda RX4 6 0 1 34.03566 34.408602 27.84810 34.150613 31.699940
#> Mazda RX4 Wag 6 0 1 34.03566 34.408602 27.84810 34.150613 34.785239
#> Datsun 710 4 1 1 11.48036 17.181037 16.48936 13.257576 16.340330
#> Hornet 4 Drive 6 1 0 27.97386 31.532633 23.86117 22.514620 23.718185
#> Hornet Sportabout 8 0 0 10.35437 8.388871 7.51073 8.411215 6.984913
#> Valiant 6 1 0 23.66013 27.499389 22.77657 20.175439 25.525636
#> qsec gear carb
#> Mazda RX4 33.605553 30.769231 28.571429
#> Mazda RX4 Wag 34.748877 30.769231 28.571429
#> Datsun 710 14.216960 13.793103 10.000000
#> Hornet 4 Drive 25.292740 21.428571 10.000000
#> Hornet Sportabout 8.273783 8.333333 5.405405
#> Valiant 26.307572 21.428571 10.000000
#> [ reached 'max' / getOption("max.print") -- omitted 26 rows ]
#>
#> Grouped by: cyl, vs, am [7 | 5 (3.8) 1-12]
mtcars |> fgroup_by(cyl,vs,am) |> fselect(mpg) |> fsum()
#> cyl vs am mpg
#> 1 4 0 1 26.0
#> 2 4 1 0 68.7
#> 3 4 1 1 198.6
#> 4 6 0 1 61.7
#> 5 6 1 0 76.5
#> 6 8 0 0 180.6
#> 7 8 0 1 30.8
## This compares fsum with data.table and base::rowsum
# Starting with small data
library(data.table)
#>
#> Attaching package: ‘data.table’
#> The following objects are masked from ‘package:dplyr’:
#>
#> between, first, last
#> The following object is masked from ‘package:collapse’:
#>
#> fdroplevels
opts <- set_collapse(nthreads = getDTthreads())
mtcDT <- qDT(mtcars)
f <- qF(mtcars$cyl)
library(microbenchmark)
microbenchmark(mtcDT[, lapply(.SD, sum), by = f],
rowsum(mtcDT, f, reorder = FALSE),
fsum(mtcDT, f, na.rm = FALSE), unit = "relative")
#> Unit: relative
#> expr min lq mean median
#> mtcDT[, lapply(.SD, sum), by = f] 77.144231 68.122951 57.312522 61.020690
#> rowsum(mtcDT, f, reorder = FALSE) 3.798077 3.659836 3.535506 3.337931
#> fsum(mtcDT, f, na.rm = FALSE) 1.000000 1.000000 1.000000 1.000000
#> uq max neval
#> 54.341880 76.14420 100
#> 3.327635 10.02351 100
#> 1.000000 1.00000 100
# Now larger data
tdata <- qDT(replicate(100, rnorm(1e5), simplify = FALSE)) # 100 columns with 100.000 obs
f <- qF(sample.int(1e4, 1e5, TRUE)) # A factor with 10.000 groups
microbenchmark(tdata[, lapply(.SD, sum), by = f],
rowsum(tdata, f, reorder = FALSE),
fsum(tdata, f, na.rm = FALSE), unit = "relative")
#> Unit: relative
#> expr min lq mean median uq
#> tdata[, lapply(.SD, sum), by = f] 3.283400 3.268921 3.232478 3.236702 3.193226
#> rowsum(tdata, f, reorder = FALSE) 2.685268 2.703754 2.522580 2.604716 2.644927
#> fsum(tdata, f, na.rm = FALSE) 1.000000 1.000000 1.000000 1.000000 1.000000
#> max neval
#> 3.941308 100
#> 1.366804 100
#> 1.000000 100
# Reset options
set_collapse(opts)