Normally, R processes data column-by-column. The data step allows you
to process data row-by-row. Row-by-row processing of data is useful when
you have related columns, and wish to perform conditional logic on those
columns. The datastep() function allows you to realize this
style of data processing. It is particularly advantageous when you wish
to perform deeply nested conditional logic. It is also very useful for
by-group processing.
Here is an example of a simple data step:
library(libr)
# Add some columns to mtcars using data step logic
df <- datastep(mtcars[1:10, 1:3], {
if (mpg >= 20)
mpgcat <- "High"
else
mpgcat <- "Low"
recdt <- as.Date("1974-06-10")
if (cyl == 8)
is8cyl <- TRUE
})
# View results
df
# mpg cyl disp recdt mpgcat is8cyl
# Mazda RX4 21.0 6 160.0 1974-06-10 High NA
# Mazda RX4 Wag 21.0 6 160.0 1974-06-10 High NA
# Datsun 710 22.8 4 108.0 1974-06-10 High NA
# Hornet 4 Drive 21.4 6 258.0 1974-06-10 High NA
# Hornet Sportabout 18.7 8 360.0 1974-06-10 Low TRUE
# Valiant 18.1 6 225.0 1974-06-10 Low NA
# Duster 360 14.3 8 360.0 1974-06-10 Low TRUE
# Merc 240D 24.4 4 146.7 1974-06-10 High NA
# Merc 230 22.8 4 140.8 1974-06-10 High NA
# Merc 280 19.2 6 167.6 1974-06-10 Low NAThe data step has parameters to perform basic shaping of the
resulting data frame. These parameters are ‘keep’, ‘drop’, and ‘rename’.
For example, the above data step could have been performed by sending
all columns into the data step, and keeping only the desired columns.
Using the keep parameter also allows you to order the
resulting columns.
library(libr)
# Keep and order output columns
df <- datastep(mtcars[1:10,],
keep = c("mpg", "cyl", "disp", "mpgcat", "recdt"), {
if (mpg >= 20)
mpgcat <- "High"
else
mpgcat <- "Low"
recdt <- as.Date("1974-06-10")
if (cyl == 8)
is8cyl <- TRUE
})
df
# mpg cyl disp mpgcat recdt
# Mazda RX4 21.0 6 160.0 High 1974-06-10
# Mazda RX4 Wag 21.0 6 160.0 High 1974-06-10
# Datsun 710 22.8 4 108.0 High 1974-06-10
# Hornet 4 Drive 21.4 6 258.0 High 1974-06-10
# Hornet Sportabout 18.7 8 360.0 Low 1974-06-10
# Valiant 18.1 6 225.0 Low 1974-06-10
# Duster 360 14.3 8 360.0 Low 1974-06-10
# Merc 240D 24.4 4 146.7 High 1974-06-10
# Merc 230 22.8 4 140.8 High 1974-06-10
# Merc 280 19.2 6 167.6 Low 1974-06-10The retain parameter allows you to define variables that will be seeded with the value from the previous step. The retain option is useful for creating cumulative values or for performing conditions based on the value of the previous row.
library(libr)
df <- datastep(mtcars[1:10, ],
drop = c("disp", "hp", "drat", "qsec",
"vs", "am", "gear", "carb"),
retain = list(cumwt = 0 ),
rename = c(mpg = "MPG", cyl = "Cylinders", wt = "Wgt",
cumwt = "Cumulative Wgt"), {
cumwt <- cumwt + wt
})
df
# MPG Cylinders Wgt Cumulative Wgt
# Mazda RX4 21.0 6 2.620 2.620
# Mazda RX4 Wag 21.0 6 2.875 5.495
# Datsun 710 22.8 4 2.320 7.815
# Hornet 4 Drive 21.4 6 3.215 11.030
# Hornet Sportabout 18.7 8 3.440 14.470
# Valiant 18.1 6 3.460 17.930
# Duster 360 14.3 8 3.570 21.500
# Merc 240D 24.4 4 3.190 24.690
# Merc 230 22.8 4 3.150 27.840
# Merc 280 19.2 6 3.440 31.280The datastep() function also has the capabilities of
performing by-group processing. A by-group is accomplished using the
by parameter, and passing a vector of column names that
define the group. Once a by-group is defined, the first.
and last. automatic variables become active, which allow
you to identify the boundaries between groups. Note that, by default,
your data must be sorted properly before sending it into the data step.
To turn the sort check off, set the sort_check parameter to
FALSE.
library(libr)
# Identify start and end of by-groups
df <- datastep(mtcars[1:10,],
keep = c("mpg", "cyl", "gear", "grp"),
by = c("gear"), sort_check = FALSE, {
if (first. & last.)
grp <- "Start - End"
else if (first.)
grp <- "Start"
else if (last.)
grp <- "End"
else
grp <- "-"
})
df
# mpg cyl gear grp
# Mazda RX4 21.0 6 4 Start
# Mazda RX4 Wag 21.0 6 4 -
# Datsun 710 22.8 4 4 End
# Hornet 4 Drive 21.4 6 3 Start
# Hornet Sportabout 18.7 8 3 -
# Valiant 18.1 6 3 -
# Duster 360 14.3 8 3 End
# Merc 240D 24.4 4 4 Start
# Merc 230 22.8 4 4 -
# Merc 280 19.2 6 4 EndIf desired, you can pass multiple variables on the by
parameter. When there are multiple by groups, the first.
and last. automatic variables described above will
represent an “or” combination of values for all by-variables. In
addition, automatic variables will be created for each variable in the
by group, similar to SAS®. Observe:
library(libr)
# Create sample data
df <- data.frame(HairEyeColor)[seq(2, 32, 2), ]
# Sort by groups
df <- sort(df, by = c("Sex", "Hair"))
# Identify start and end of by-groups
df2 <- datastep(df,
drop = c("Eye", "Freq"),
by = c("Sex", "Hair"), {
fSex <- first.Sex
lSex <- last.Sex
fHair <- first.Hair
lHair <- last.Hair
})
df2
# Hair Sex fSex lSex fHair lHair
# 1 Brown Male TRUE FALSE TRUE FALSE
# 2 Brown Male FALSE FALSE FALSE FALSE
# 3 Brown Male FALSE FALSE FALSE FALSE
# 4 Brown Male FALSE FALSE FALSE TRUE
# 5 Blond Male FALSE FALSE TRUE FALSE
# 6 Blond Male FALSE FALSE FALSE FALSE
# 7 Blond Male FALSE FALSE FALSE FALSE
# 8 Blond Male FALSE TRUE FALSE TRUE
# 9 Brown Female TRUE FALSE TRUE FALSE
# 10 Brown Female FALSE FALSE FALSE FALSE
# 11 Brown Female FALSE FALSE FALSE FALSE
# 12 Brown Female FALSE FALSE FALSE TRUE
# 13 Blond Female FALSE FALSE TRUE FALSE
# 14 Blond Female FALSE FALSE FALSE FALSE
# 15 Blond Female FALSE FALSE FALSE FALSE
# 16 Blond Female FALSE TRUE FALSE TRUEThe above first.Sex, last.Sex,
first.Hair, and last.Hair variables may also
be used in conditions, functions, or any other expression inside your
datastep. Note that like first. and last. they
are dropped automatically at the end of the datastep. If you want to
retain their values, assign them to a new variable as shown above.
There may be times when you want to combine row-by-row conditional
processing with column-by-column vector operations. For example, let’s
say you want to calculate a mean and then perform conditional processing
on that mean. This situation can be handled using the
calculate parameter on the datastep()
function. The function will execute the calculate block
first, add any assigned variables to the data frame, and then execute
the data step. Below is an example of such a scenario:
library(libr)
# Categorize mpg as above or below the mean
df <- datastep(mtcars,
keep = c("mpg", "cyl", "mean_mpg", "mpgcat"),
calculate = { mean_mpg = mean(mpg) },
{
if (mpg >= mean_mpg)
mpgcat <- "High"
else
mpgcat <- "Low"
})
df[1:10,]
# mpg cyl mean_mpg mpgcat
# Mazda RX4 21.0 6 20.09062 High
# Mazda RX4 Wag 21.0 6 20.09062 High
# Datsun 710 22.8 4 20.09062 High
# Hornet 4 Drive 21.4 6 20.09062 High
# Hornet Sportabout 18.7 8 20.09062 Low
# Valiant 18.1 6 20.09062 Low
# Duster 360 14.3 8 20.09062 Low
# Merc 240D 24.4 4 20.09062 High
# Merc 230 22.8 4 20.09062 High
# Merc 280 19.2 6 20.09062 LowdplyrNote that the datastep() function is pipe-friendly, and
can be combined with dplyr functions in a data
pipeline. Also note that the datastep() function will
recognize any group attributes added by the group_by()
function. Therefore, within a dplyr pipeline, it is not
necessary to use any datastep parameters. The following
example recreates the above data frame from Example 5, but with a
dplyr pipeline.
library(libr)
library(dplyr)
library(magrittr)
# Add datastep to dplyr pipeline
df <- mtcars %>%
select(mpg, cyl, gear) %>%
mutate(mean_mpg = mean(mpg)) %>%
datastep({
if (mpg >= mean_mpg)
mpgcat <- "High"
else
mpgcat <- "Low"
}) %>%
filter(row_number() <= 10)
df
# mpg cyl gear mean_mpg mpgcat
# 1 21.0 6 4 20.09062 High
# 2 21.0 6 4 20.09062 High
# 3 22.8 4 4 20.09062 High
# 4 21.4 6 3 20.09062 High
# 5 18.7 8 3 20.09062 Low
# 6 18.1 6 3 20.09062 Low
# 7 14.3 8 3 20.09062 Low
# 8 24.4 4 4 20.09062 High
# 9 22.8 4 4 20.09062 High
# 10 19.2 6 4 20.09062 LowThe libr package recognizes several useful data attributes that are not normally recognized by other R functions. For example, it is very convenient to assign label and description attributes to your columns, so other people can understand what data the columns contain.
For this reason, the datastep() function provides an
attrib parameter that allows you to supply such attributes as
part of a data step. Attributes are assigned with a named list and the
dsattr() object.
library(libr)
# Assign label attributes to all columns
df <- datastep(mtcars[1:10, ],
keep = c("mpg", "cyl", "mpgcat"),
calculate = { mean_mpg = mean(mpg) },
attrib = list(mpg = dsattr(label = "Miles Per Gallon"),
cyl = dsattr(label = "Cylinders"),
mpgcat = dsattr(label = "Mileage Category")), {
if (mpg >= mean_mpg)
mpgcat <- "High"
else
mpgcat <- "Low"
})
# View attributes in dictionary
dictionary(df)
# # A tibble: 3 x 10
# Name Column Class Label Description Format Width Justify Rows NAs
# <chr> <chr> <chr> <chr> <chr> <lgl> <int> <chr> <int> <int>
# 1 df mpg numeric Miles Per Gallon NA NA NA NA 10 0
# 2 df cyl numeric Cylinders NA NA NA NA 10 0
# 3 df mpgcat character Mileage Category NA NA 4 NA 10 0As mentioned previously, R typically operates in a column-wise manner. That is, R processes data column-by-column. But what if you need to get a sum or mean across a row?
This situation is what led to the development of the data
step array. The data step array allows you to define a list of
columns and
iterate over the list inside a data step. Data step arrays are defined
with the arrays parameter, which accepts a named list of
dsarray() objects.
To see the array in action, we’ll use the AirPassengers sample data. This data shows international airline passengers by month between 1949 and 1960. The data looks like this:
AirPassengers
# Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
# 1949 112 118 132 129 121 135 148 148 136 119 104 118
# 1950 115 126 141 135 125 149 170 170 158 133 114 140
# 1951 145 150 178 163 172 178 199 199 184 162 146 166
# 1952 171 180 193 181 183 218 230 242 209 191 172 194
# 1953 196 196 236 235 229 243 264 272 237 211 180 201
# 1954 204 188 235 227 234 264 302 293 259 229 203 229
# 1955 242 233 267 269 270 315 364 347 312 274 237 278
# 1956 284 277 317 313 318 374 413 405 355 306 271 306
# 1957 315 301 356 348 355 422 465 467 404 347 305 336
# 1958 340 318 362 348 363 435 491 505 404 359 310 337
# 1959 360 342 406 396 420 472 548 559 463 407 362 405
# 1960 417 391 419 461 472 535 622 606 508 461 390 432This example illustrates how to create row totals, row means, and find the top month using a data step array. The array has an indexer to extract values. You can use the indexer to extract a single value or a subset of values. An empty indexer will return all the values in the array.
library(libr)
# Create AirPassengers Data Frame
df <- as.data.frame(t(matrix(AirPassengers, 12,
dimnames = list(month.abb, seq(1949, 1960)))),
stringsAsFactors = FALSE)
# Use datastep array to get year tot, mean, and top month
dat <- datastep(df,
arrays = list(months = dsarray(names(df))),
attrib = list(Tot = 0, Mean = 0, Top = ""),
drop = "mth",
{
Tot <- sum(months[])
Mean <- mean(months[])
for (mth in months) {
if (months[mth] == max(months[])) {
Top <- mth
}
}
})
dat
# Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Tot Mean Top
# 1949 112 118 132 129 121 135 148 148 136 119 104 118 1520 126.6667 Aug
# 1950 115 126 141 135 125 149 170 170 158 133 114 140 1676 139.6667 Aug
# 1951 145 150 178 163 172 178 199 199 184 162 146 166 2042 170.1667 Aug
# 1952 171 180 193 181 183 218 230 242 209 191 172 194 2364 197.0000 Aug
# 1953 196 196 236 235 229 243 264 272 237 211 180 201 2700 225.0000 Aug
# 1954 204 188 235 227 234 264 302 293 259 229 203 229 2867 238.9167 Jul
# 1955 242 233 267 269 270 315 364 347 312 274 237 278 3408 284.0000 Jul
# 1956 284 277 317 313 318 374 413 405 355 306 271 306 3939 328.2500 Jul
# 1957 315 301 356 348 355 422 465 467 404 347 305 336 4421 368.4167 Aug
# 1958 340 318 362 348 363 435 491 505 404 359 310 337 4572 381.0000 Aug
# 1959 360 342 406 396 420 472 548 559 463 407 362 405 5140 428.3333 Aug
# 1960 417 391 419 461 472 535 622 606 508 461 390 432 5714 476.1667 JulIn the example above, the “Tot”, “Mean”, and “Top” columns were all calculated using the datastep array. These types of row-wise statistics are hard to calculate otherwise.
The datastep provides different ways to control which rows are output.
First, the function has a where parameter to pass a
filter expression to the datastep. The where clause will be executed at
the end of datastep processing. Pass in the where clause using the
expression() function. Like so:
# Prepare sample data
dat <- as.data.frame(HairEyeColor)
# Filter for black hair and blue eyes
res <- datastep(dat,
where = expression(Hair == "Black" & Eye == "Blue"),
{})
res
# Hair Eye Sex Freq
# 1 Black Blue Male 11
# 2 Black Blue Female 9The datastep also recognizes the delete() and
output() functions to remove or duplicate rows from inside
the datastep. These functions give you conditional control over which
rows are output.
# Delete rows with frequencies less than 25
res1 <- datastep(dat, {
if (Freq < 25)
delete()
})
res1
# Hair Eye Sex Freq
# 1 Black Brown Male 32
# 2 Brown Brown Male 53
# 3 Brown Blue Male 50
# 4 Blond Blue Male 30
# 5 Brown Hazel Male 25
# 6 Black Brown Female 36
# 7 Brown Brown Female 66
# 8 Brown Blue Female 34
# 9 Blond Blue Female 64
# 10 Brown Hazel Female 29
# Only output rows for brown-eyes and frequencies over 25
res2 <- datastep(dat, {
if (Eye == "Brown") {
if (Freq >= 25) {
output()
}
}
})
res2
# Hair Eye Sex Freq
# 1 Black Brown Male 32
# 2 Brown Brown Male 53
# 3 Black Brown Female 36
# 4 Brown Brown Female 66You can also use the output() function to create
datasets from scratch, just like in SAS®. To create a dataset from
scratch, simply pass in an empty data frame and output the desired
values.
# Create metadata
res3 <- datastep(data.frame(), {
name <- "mtcars"
rows <- nrow(mtcars)
cols <- ncol(mtcars)
output()
name <- "iris"
rows <- nrow(iris)
cols <- ncol(iris)
output()
name <- "beaver1"
rows <- nrow(beaver1)
cols <- ncol(beaver1)
output()
})
res3
# name rows cols
# 1 mtcars 32 11
# 2 iris 150 5
# 3 beaver1 114 4When working with data, joining datasets is an essential activity.
While there are many different functions in R to perform joins, the
datastep() “set” and “merge” parameters offer unusual
flexibility.
The “set” parameter stacks two or more datasets. The “merge” parameter joins two or more datasets. Together, these two parameters allow you to perform the most common types of data combinations.
To illustrate, first let’s create up some sample datasets. The datasets we will create include one “region” dataset, and two “stores” datasets. Note that the columns on the stores datasets are not identical.
# Create sample data
region <- read.table(header = TRUE, text = '
REGION NAME
R01 East
R02 West
R03 North
R04 South
', stringsAsFactors = FALSE)
# First stores dataset
stores1 <- read.table(header = TRUE, text = '
ID NAME SIZE REGION FRANCHISE
A01 "Eastern Lumber" L R01 T
A02 "Tri-City Hardwood" M R02 F
A05 "Reliable Hardware" S R01 T
', stringsAsFactors = FALSE)
# Extra column on this one
stores2 <- read.table(header = TRUE, text = '
ID NAME SIZE REGION
A03 "AAA Mills" S R05
A04 "Home and Yard" L R03
', stringsAsFactors = FALSE)Despite not having the same columns, the two stores datasets can be
set using the datastep() function. The function will fill
in the missing values automatically. Like so:
# Set operation
allstores <- datastep(stores1, set = stores2, {})
# Extra values filled with NA
allstores
# ID NAME SIZE REGION FRANCHISE
# 1 A01 Eastern Lumber L R01 TRUE
# 2 A02 Tri-City Hardwood M R02 FALSE
# 3 A05 Reliable Hardware S R01 TRUE
# 4 A03 AAA Mills S R05 NA
# 5 A04 Home and Yard L R03 NALet’s pretend we noticed the missing data, and decide to fill it in. We can do that by merging the missing FRANCHISE values to the second stores dataset. First let’s create the missing data:
# Create small dataset of missing FRANCHISE values
franchises <- data.frame(FRANCHISE = c(F, F), stringsAsFactors = FALSE)
franchises
# FRANCHISE
# 1 FALSE
# 2 FALSENext we can merge in the missing data on “stores2”, and set the two store datasets again:
# Merge in missing FRANCHISE column
stores2mod <- datastep(stores2, merge = franchises, {})
stores2mod
# ID NAME SIZE REGION FRANCHISE
# 1 A03 AAA Mills S R05 FALSE
# 2 A04 Home and Yard L R03 FALSE
# Set again
allstores <- datastep(stores1, set = stores2mod, {})
# Now everything is aligned
allstores
# ID NAME SIZE REGION FRANCHISE
# 1 A01 Eastern Lumber L R01 TRUE
# 2 A02 Tri-City Hardwood M R02 FALSE
# 3 A05 Reliable Hardware S R01 TRUE
# 4 A03 AAA Mills S R05 FALSE
# 5 A04 Home and Yard L R03 FALSEObserve that we did not have to specify a join condition on the merge. When no “merge_by” is indicated, the datastep will simply append the new columns to the right - without complaining. This behavior is very convenient.
Now let’s do another join, but this time we will specify a join condition. We will join in the store regions by the region ID. We will also set up merge flags so we can see which rows were in which input dataset.
# Merge operation - Outer Join
res <- datastep(allstores, merge = region,
merge_by = "REGION",
merge_in = c("inA", "inB"), {})
# View results
res
# ID NAME.1 SIZE REGION FRANCHISE NAME.2 inA inB
# 1 A01 Eastern Lumber L R01 TRUE East 1 1
# 2 A05 Reliable Hardware S R01 TRUE East 1 1
# 3 A02 Tri-City Hardwood M R02 FALSE West 1 1
# 4 A04 Home and Yard L R03 FALSE North 1 1
# 5 A03 AAA Mills S R05 FALSE <NA> 1 0
# 6 <NA> <NA> <NA> R04 NA South 0 1Notice three things:
Let’s try one last time to fix the above issues. We can fix the column names with the “rename” parameter, and exclude rows with a “where” expression. Also, we can recode “R05” to “R04” inside the datastep. Finally, we’ll drop the merge flags to clean up the columns.
# Merge operation - Left join and clean up
res <- datastep(allstores, merge = region,
merge_by = "REGION",
merge_in = c("inA", "inB"),
rename = c(NAME.1 = "STORE_NAME", NAME.2 = "REGION_NAME"),
where = expression(inA == TRUE),
drop = c("inA", "inB"),
{
if (REGION == "R05") {
REGION <- "R04"
NAME.2 <- "South"
}
})
#'
# View results
res
# ID STORE_NAME SIZE REGION FRANCHISE REGION_NAME
# 1 A01 Eastern Lumber L R01 TRUE East
# 2 A05 Reliable Hardware S R01 TRUE East
# 3 A02 Tri-City Hardwood M R02 FALSE West
# 4 A04 Home and Yard L R03 FALSE North
# 5 A03 AAA Mills S R04 FALSE SouthOne weakness of the libr datastep()
function is performance. The function is far slower than the equivalent
SAS® datastep. The performance profile may limit the number of records
you are able to reasonably process with the datastep().
One thing you can do to increase performance is to reduce the number of rows and columns on the input data. You can perform this pre-filtering with Base R or Tidyverse functions. This strategy is particularly recommended if you were planning to subset the data anyway using the “where” or “keep” options.
The Base R subset() function is convenient to use
because it is always available. Here is an example showing how to reduce
the size of the iris sample dataframe using Base R
subset() before sending it to a datastep.
# Subset the input dataset first for only needed rows and columns
dat <- subset(iris, Species == 'versicolor', c('Petal.Length', 'Petal.Width')) |>
datastep({
if (Petal.Length < 3.5)
Petal.Size <- "Short"
else if (Petal.Length > 4.5)
Petal.Size <- "Long"
else
Petal.Size <- "Medium"
})
# View Some Results
dat[1:10, ]
# Petal.Length Petal.Width Petal.Size
# 1 4.7 1.4 Long
# 2 4.5 1.5 Medium
# 3 4.9 1.5 Long
# 4 4.0 1.3 Medium
# 5 4.6 1.5 Long
# 6 4.5 1.3 Medium
# 7 4.7 1.6 Long
# 8 3.3 1.0 Short
# 9 4.6 1.3 Long
# 10 3.9 1.4 Medium
Next: Disclaimer