An introduction to data.table

R
beginner
Published

August 13, 2025

No feedback found for this session

Introduction

data.table is an enhanced data.frame. It’s fast and concise, and feels very like an improved version of base-R tools for working with tabular data. This training session gives a compare-and-contrast introduction to data.table by running through some core data operations in data.table and by comparing those to similar operations in base-R and tidyverse.

There’s some sample data that you might like to download and place in your project.

There’s also a bit of benchmark data here. This was generated in R4.4.1, Rstudio desktop 2025.05.1 on Windows 10, using a fairly new-ish laptop (Intel i7-13850HX, 128Gb RAM).

Base R data.frames

In base R, the usual way of holding rectangular data is using a data frame. Perhaps the best-known example is mtcars:

mtcars
              mpg cyl disp  hp drat    wt  qsec vs am gear carb
Mazda RX4      21   6  160 110  3.9 2.620 16.46  0  1    4    4
Mazda RX4 Wag  21   6  160 110  3.9 2.875 17.02  0  1    4    4
 [ reached 'max' / getOption("max.print") -- omitted 30 rows ]

As a reminder, there are several techniques for manipulating that data frame. For example, we can subset the rows/cols of mtcars to discard parts of the data:

mtcars[1] # column by index
                     mpg
Mazda RX4           21.0
Mazda RX4 Wag       21.0
Datsun 710          22.8
Hornet 4 Drive      21.4
Hornet Sportabout   18.7
Valiant             18.1
Duster 360          14.3
Merc 240D           24.4
Merc 230            22.8
Merc 280            19.2
Merc 280C           17.8
Merc 450SE          16.4
Merc 450SL          17.3
Merc 450SLC         15.2
Cadillac Fleetwood  10.4
Lincoln Continental 10.4
Chrysler Imperial   14.7
Fiat 128            32.4
Honda Civic         30.4
Toyota Corolla      33.9
Toyota Corona       21.5
Dodge Challenger    15.5
AMC Javelin         15.2
Camaro Z28          13.3
Pontiac Firebird    19.2
 [ reached 'max' / getOption("max.print") -- omitted 7 rows ]
mtcars["mpg"] # column by name
                     mpg
Mazda RX4           21.0
Mazda RX4 Wag       21.0
Datsun 710          22.8
Hornet 4 Drive      21.4
Hornet Sportabout   18.7
Valiant             18.1
Duster 360          14.3
Merc 240D           24.4
Merc 230            22.8
Merc 280            19.2
Merc 280C           17.8
Merc 450SE          16.4
Merc 450SL          17.3
Merc 450SLC         15.2
Cadillac Fleetwood  10.4
Lincoln Continental 10.4
Chrysler Imperial   14.7
Fiat 128            32.4
Honda Civic         30.4
Toyota Corolla      33.9
Toyota Corona       21.5
Dodge Challenger    15.5
AMC Javelin         15.2
Camaro Z28          13.3
Pontiac Firebird    19.2
 [ reached 'max' / getOption("max.print") -- omitted 7 rows ]
mtcars[1,] # row by index
          mpg cyl disp  hp drat   wt  qsec vs am gear carb
Mazda RX4  21   6  160 110  3.9 2.62 16.46  0  1    4    4
mtcars[1,][1] # value by row and column index
          mpg
Mazda RX4  21

There are also several ways of extracting parts of that data as a vector:

mtcars[,1] # column by index
 [1] 21.0 21.0 22.8 21.4 18.7 18.1 14.3 24.4 22.8 19.2 17.8 16.4 17.3 15.2 10.4
[16] 10.4 14.7 32.4 30.4 33.9 21.5 15.5 15.2 13.3 19.2
 [ reached 'max' / getOption("max.print") -- omitted 7 entries ]
mtcars[["mpg"]] # column by name
 [1] 21.0 21.0 22.8 21.4 18.7 18.1 14.3 24.4 22.8 19.2 17.8 16.4 17.3 15.2 10.4
[16] 10.4 14.7 32.4 30.4 33.9 21.5 15.5 15.2 13.3 19.2
 [ reached 'max' / getOption("max.print") -- omitted 7 entries ]
mtcars$mpg # alternative, sweetened, way of column by name
 [1] 21.0 21.0 22.8 21.4 18.7 18.1 14.3 24.4 22.8 19.2 17.8 16.4 17.3 15.2 10.4
[16] 10.4 14.7 32.4 30.4 33.9 21.5 15.5 15.2 13.3 19.2
 [ reached 'max' / getOption("max.print") -- omitted 7 entries ]
mtcars[1,] |> 
  unlist() # usual way of taking a row as a vector
   mpg    cyl   disp     hp   drat     wt   qsec     vs     am   gear   carb 
 21.00   6.00 160.00 110.00   3.90   2.62  16.46   0.00   1.00   4.00   4.00 
mtcars[3,7] # single value by row and column index
[1] 18.61

Much of the data.table syntax we’ll look at in this session has a passing resemblance to these base R ways of working, but with several important differences. It’s therefore useful to (re)familiarise yourself with these basics before moving on to the new material.

Package loading

Reading data

This session is based on comparisons of simple tasks between base R, tidyverse, and data.table. We’ll start with three different methods for loading csv data which create subtly different objects in the three cases.

df <- read.csv(here(local_path))
tb <- read_csv(here(local_path))
dt <- fread(here(local_path))

In case these violin plots are new to you: these are the output of the microbenchmark package, which repeatedly runs bits of code while timing their speed of execution. The wider the violin (the white blobs) the greater the number of times the code ran at that speed. The further left the blob, the faster the code. Here, data.table’s fread ran much much faster than either base-R read.csv or tidyverse read_csv. There is some code for the benchmarking data here which is included for interest.

Code
# data reading ----
dat_read <- microbenchmark(
  df <- read.csv(local_path),
  tb <- read_csv(local_path),
  dt <- fread(local_path)
  )

# col to col ----

col_subset <- microbenchmark(
  df["month"], # base-R ish subsetting
  tb |>
    select(month), # tidy
  dt[,.(month)] # dt
  )

# col to vec ----

vec_subset2 <- microbenchmark(
  df$month, # base-R ish subsetting
  tb |>
    pull(month), # tidy
  dt[, month] # usual DT approach
  )

# rows ----

row_subset <- microbenchmark(
  df[1:2,], # rows
  tb |>
    slice(1:2),
  dt[1:2]
  )

# filter ----

filter_subset <- microbenchmark(
  df[(df$dest == "LAX"),],
  tb |>
    filter(dest == "LAX"),
  dt[dest == "LAX"]
  )

# ungrouped summarize ----

ungrouped_summaries <- microbenchmark(
  data.frame(mean = mean(df$hour)),

  tb |>
    summarise(mean = mean(hour)),

  dt[,.(mean = mean(hour))]
  )

# grouped summarize ----

grouped_summaries <- microbenchmark(
  aggregate(hour ~ carrier, data = df, FUN = mean),

  tb |>
    group_by(carrier) |>
    summarise(mean(hour)),

  dt[,mean(hour), carrier]
  )

# subset then group then summarise ----

subset_grouped_summaries <- microbenchmark(
  aggregate(distance ~ carrier, data = subset(df, month == 9), FUN = "sum"),
  
  tb |>
    filter(month == 9) |>
    group_by(carrier) |>
    summarise(sum = sum(distance)),
  
  dt[month == 9, sum(distance), carrier]
)

# arrange ----

arranged <- microbenchmark(
  df[order(df$dep_delay),],
  tb |>
    arrange(dep_delay),
  dt[order(dep_delay)]
)
autoplot(arranged)

# count ----

counts <- microbenchmark(
  data.frame(table(df$carrier)),
  tb |>
    count(carrier),
  dt[, .(n = .N), by = carrier]
)

# mutate ----

mutates <- microbenchmark(
  {df$dist_km <- df$distance * 1.6
  df[c("dist_km", "carrier", "origin", "dest")]},
  tidyverse = tb |>
    mutate(dist_km = distance * 1.6) |>
    select(dist_km, carrier, origin, dest),
  dt[,.(dist_km = distance * 1.6, carrier, origin, dest)],
  dt[,.(dist_km = distance*1.6),.(carrier,origin, dest)]
  
)

# subset several cols

col_subset_multiple <- microbenchmark(
  df[c("day", "month", "year")],
  tb |>
    select(day, month, year),
  dt[, list("day", "month", "year")]
)

bench_dat <- list(dat_read = dat_read, 
                  col_subset = col_subset, 
                  vec_subset = vec_subset, 
                  row_subset = row_subset, 
                  filter_subset = filter_subset, 
                  ungrouped_summaries = ungrouped_summaries, 
                  grouped_summaries = grouped_summaries, 
                  subset_grouped_summaries = subset_grouped_summaries, 
                  arranged = arranged, 
                  counts = counts, 
                  mutates = mutates,
                  col_subset_multiple = col_subset_multiple) 
 
write_rds(bench_dat, "data/bench_dat.rds")

Objects

Note that the three different csv reading functions create different slightly different objects:

class(tb)
[1] "spec_tbl_df" "tbl_df"      "tbl"         "data.frame" 
class(dt)
[1] "data.table" "data.frame"
class(df)
[1] "data.frame"

data.table has its own data structure: obviously enough, the data.table. You can create data.tables using fread, or with care, convert data.frames/tibbles using setDT. The reason for needing care is that setDT (and a few other functions in data.table) modify in place. This is one of the few cases in R where you can change an object without assigning:

test <- tibble(alph = letters) # make a tibble
class(test) # shows "tbl_df"     "tbl"
[1] "tbl_df"     "tbl"        "data.frame"
setDT(test) # note modify-in-place
class(test) # that's a data.table
[1] "data.table" "data.frame"
setDF(test) # back into a df
class(test) # and that's a data.frame
[1] "data.frame"
test <- as_tibble(test) # if you need a tibble
class(test)
[1] "tbl_df"     "tbl"        "data.frame"

What’s the big idea

If you’re used to working with tidyverse, data.table takes a radically different approach to working with data. The standard way of representing data.table syntax is as follows:

dt[i, j, by]

This means:

  • take dt - a data.table
  • subset or reorder its rows by i
  • subset columns or calculate by j
  • while grouping by by

Unless you’ve done lots of SQL before, that’s likely to be confusing. So let’s do some simple operations first, then return to the overview.

Subset one column to single-column df/tibble/dt

df["month"]
   month
1      1
2      1
3      1
4      1
5      1
6      1
7      1
8      1
9      1
10     1
11     1
12     1
13     1
14     1
15     1
16     1
17     1
18     1
19     1
20     1
21     1
22     1
23     1
24     1
25     1
 [ reached 'max' / getOption("max.print") -- omitted 253291 rows ]
tb |> 
  select(month)
# A tibble: 253,316 × 1
   month
   <dbl>
 1     1
 2     1
 3     1
 4     1
 5     1
 6     1
 7     1
 8     1
 9     1
10     1
# ℹ 253,306 more rows
dt[, "month"]
        month
        <int>
     1:     1
     2:     1
     3:     1
     4:     1
     5:     1
    ---      
253312:    10
253313:    10
253314:    10
253315:    10
253316:    10
# or
dt[, list(month)]
        month
        <int>
     1:     1
     2:     1
     3:     1
     4:     1
     5:     1
    ---      
253312:    10
253313:    10
253314:    10
253315:    10
253316:    10
# or
dt[, .(month)]
        month
        <int>
     1:     1
     2:     1
     3:     1
     4:     1
     5:     1
    ---      
253312:    10
253313:    10
253314:    10
253315:    10
253316:    10

Subset several columns to df/tibble/dt

df[c("day", "month", "year")]
  day month year
1   1     1 2014
2   1     1 2014
3   1     1 2014
4   1     1 2014
5   1     1 2014
6   1     1 2014
7   1     1 2014
8   1     1 2014
 [ reached 'max' / getOption("max.print") -- omitted 253308 rows ]
tb |>
  select(day, month, year)
# A tibble: 253,316 × 3
     day month  year
   <dbl> <dbl> <dbl>
 1     1     1  2014
 2     1     1  2014
 3     1     1  2014
 4     1     1  2014
 5     1     1  2014
 6     1     1  2014
 7     1     1  2014
 8     1     1  2014
 9     1     1  2014
10     1     1  2014
# ℹ 253,306 more rows
dt[, list(day, month, year)] # list
          day month  year
        <int> <int> <int>
     1:     1     1  2014
     2:     1     1  2014
     3:     1     1  2014
     4:     1     1  2014
     5:     1     1  2014
    ---                  
253312:    31    10  2014
 [ reached 'max' / getOption("max.print") -- omitted 4 rows ]
dt[,.(day, month, year)] # idiomatic list shorthand
          day month  year
        <int> <int> <int>
     1:     1     1  2014
     2:     1     1  2014
     3:     1     1  2014
     4:     1     1  2014
     5:     1     1  2014
    ---                  
253312:    31    10  2014
 [ reached 'max' / getOption("max.print") -- omitted 4 rows ]

Single column to vector

df$month
 [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
 [ reached 'max' / getOption("max.print") -- omitted 253291 entries ]
tb |>
  pull(month)
 [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
 [ reached 'max' / getOption("max.print") -- omitted 253291 entries ]
dt[, month] 
 [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
 [ reached 'max' / getOption("max.print") -- omitted 253291 entries ]

Subset rows

df[1:2,] # rows
  year month day dep_delay arr_delay carrier origin dest air_time distance hour
1 2014     1   1        14        13      AA    JFK  LAX      359     2475    9
2 2014     1   1        -3        13      AA    JFK  LAX      363     2475   11
tb |>
  slice(1:2)
# A tibble: 2 × 11
   year month   day dep_delay arr_delay carrier origin dest  air_time distance
  <dbl> <dbl> <dbl>     <dbl>     <dbl> <chr>   <chr>  <chr>    <dbl>    <dbl>
1  2014     1     1        14        13 AA      JFK    LAX        359     2475
2  2014     1     1        -3        13 AA      JFK    LAX        363     2475
# ℹ 1 more variable: hour <dbl>
dt[1:2]
    year month   day dep_delay arr_delay carrier origin   dest air_time
   <int> <int> <int>     <int>     <int>  <char> <char> <char>    <int>
1:  2014     1     1        14        13      AA    JFK    LAX      359
   distance  hour
      <int> <int>
1:     2475     9
 [ reached 'max' / getOption("max.print") -- omitted 1 row ]

Filter rows

df[(df$dest == "LAX"),]
  year month day dep_delay arr_delay carrier origin dest air_time distance hour
1 2014     1   1        14        13      AA    JFK  LAX      359     2475    9
2 2014     1   1        -3        13      AA    JFK  LAX      363     2475   11
 [ reached 'max' / getOption("max.print") -- omitted 14432 rows ]
tb |>
  filter(dest == "LAX")
# A tibble: 14,434 × 11
    year month   day dep_delay arr_delay carrier origin dest  air_time distance
   <dbl> <dbl> <dbl>     <dbl>     <dbl> <chr>   <chr>  <chr>    <dbl>    <dbl>
 1  2014     1     1        14        13 AA      JFK    LAX        359     2475
 2  2014     1     1        -3        13 AA      JFK    LAX        363     2475
 3  2014     1     1         2         9 AA      JFK    LAX        351     2475
 4  2014     1     1         2         1 AA      JFK    LAX        350     2475
 5  2014     1     1         4         0 AA      EWR    LAX        339     2454
 6  2014     1     1        -2       -18 AA      JFK    LAX        338     2475
 7  2014     1     1        -3       -14 AA      JFK    LAX        356     2475
 8  2014     1     1       142       133 AA      JFK    LAX        345     2475
 9  2014     1     1        -4        11 B6      JFK    LAX        349     2475
10  2014     1     1         3       -10 B6      JFK    LAX        349     2475
# ℹ 14,424 more rows
# ℹ 1 more variable: hour <dbl>
dt[dest == "LAX"]
        year month   day dep_delay arr_delay carrier origin   dest air_time
       <int> <int> <int>     <int>     <int>  <char> <char> <char>    <int>
    1:  2014     1     1        14        13      AA    JFK    LAX      359
       distance  hour
          <int> <int>
    1:     2475     9
 [ reached 'max' / getOption("max.print") -- omitted 10 rows ]

Summarize

so return a new df/tibble/dt

data.frame(mean = mean(df$hour))
      mean
1 13.06343
tb |>
  summarise(mean = mean(hour))
# A tibble: 1 × 1
   mean
  <dbl>
1  13.1
dt[,.(mean = mean(hour))]
       mean
      <num>
1: 13.06343

Group and summarize

aggregate(hour ~ carrier, data = df, FUN = "mean")
   carrier      hour
1       AA 12.389590
2       AS 12.881533
3       B6 13.430518
4       DL 13.007605
5       EV 13.535121
6       F9 13.534884
7       FL 14.270184
8       HA  9.115385
9       MQ 13.750849
10      OO 10.670000
11      UA 12.997536
12      US 12.050448
 [ reached 'max' / getOption("max.print") -- omitted 2 rows ]
tb |>
  group_by(carrier) |>
  summarise(mean(hour))
# A tibble: 14 × 2
   carrier `mean(hour)`
   <chr>          <dbl>
 1 AA             12.4 
 2 AS             12.9 
 3 B6             13.4 
 4 DL             13.0 
 5 EV             13.5 
 6 F9             13.5 
 7 FL             14.3 
 8 HA              9.12
 9 MQ             13.8 
10 OO             10.7 
11 UA             13.0 
12 US             12.1 
13 VX             12.3 
14 WN             12.7 
dt[,mean(hour), carrier]
    carrier        V1
     <char>     <num>
 1:      AA 12.389590
 2:      AS 12.881533
 3:      B6 13.430518
 4:      DL 13.007605
 5:      EV 13.535121
 6:      F9 13.534884
 7:      FL 14.270184
 8:      HA  9.115385
 9:      MQ 13.750849
10:      VX 12.322493
11:      WN 12.696354
 [ reached 'max' / getOption("max.print") -- omitted 3 rows ]
autoplot(bench_dat$grouped_summaries)

Subset, group, summarise

aggregate(distance ~ carrier, data = subset(df, month == 9), FUN = "sum")
   carrier distance
1       AA  3687744
2       AS   144120
3       B6  4613124
4       DL  5094885
5       EV  2386348
6       F9    90720
7       HA   124575
8       MQ   906560
9       OO    35184
10      UA  7109454
11      US   906256
12      VX  1218951
 [ reached 'max' / getOption("max.print") -- omitted 1 rows ]
tb |>
  filter(month == 9) |>
  group_by(carrier) |>
  summarise(sum = sum(distance))
# A tibble: 13 × 2
   carrier     sum
   <chr>     <dbl>
 1 AA      3687744
 2 AS       144120
 3 B6      4613124
 4 DL      5094885
 5 EV      2386348
 6 F9        90720
 7 HA       124575
 8 MQ       906560
 9 OO        35184
10 UA      7109454
11 US       906256
12 VX      1218951
13 WN      1297176
dt[month == 9, sum(distance), carrier]
    carrier      V1
     <char>   <int>
 1:      AA 3687744
 2:      AS  144120
 3:      B6 4613124
 4:      DL 5094885
 5:      VX 1218951
 6:      WN 1297176
 7:      EV 2386348
 8:      F9   90720
 9:      HA  124575
10:      MQ  906560
11:      OO   35184
 [ reached 'max' / getOption("max.print") -- omitted 2 rows ]
autoplot(bench_dat$subset_grouped_summaries)

Arrange

df[order(df$dep_delay),]
      year month day dep_delay arr_delay carrier origin dest air_time distance
14836 2014     1  21      -112      -112      DL    LGA  MSP      153     1020
70868 2014     4   1       -34       -40      DL    JFK  SJU      192     1598
      hour
14836   14
70868   23
 [ reached 'max' / getOption("max.print") -- omitted 253314 rows ]
tb |>
  arrange(dep_delay)
# A tibble: 253,316 × 11
    year month   day dep_delay arr_delay carrier origin dest  air_time distance
   <dbl> <dbl> <dbl>     <dbl>     <dbl> <chr>   <chr>  <chr>    <dbl>    <dbl>
 1  2014     1    21      -112      -112 DL      LGA    MSP        153     1020
 2  2014     4     1       -34       -40 DL      JFK    SJU        192     1598
 3  2014     1     9       -27        18 AA      EWR    LAX        332     2454
 4  2014     2     2       -27       -42 EV      LGA    BGR         51      378
 5  2014     8    26       -25       -24 B6      LGA    FLL        134     1076
 6  2014     1    13       -24        -1 AA      EWR    DFW        229     1372
 7  2014     8    26       -24       -41 B6      EWR    MCO        123      937
 8  2014     8    31       -24       -30 B6      EWR    MCO        120      937
 9  2014    10    28       -24       -38 B6      EWR    FLL        142     1065
10  2014     1    14       -23       -24 EV      EWR    MHT         42      209
# ℹ 253,306 more rows
# ℹ 1 more variable: hour <dbl>
dt[order(dep_delay)]
         year month   day dep_delay arr_delay carrier origin   dest air_time
        <int> <int> <int>     <int>     <int>  <char> <char> <char>    <int>
     1:  2014     1    21      -112      -112      DL    LGA    MSP      153
        distance  hour
           <int> <int>
     1:     1020    14
 [ reached 'max' / getOption("max.print") -- omitted 10 rows ]
autoplot(bench_dat$arranged)

Count

data.frame(table(df$carrier))
   Var1  Freq
1    AA 26302
2    AS   574
3    B6 44479
4    DL 41683
5    EV 39819
6    F9   473
7    FL  1251
8    HA   260
9    MQ 18559
10   OO   200
11   UA 46267
12   US 16750
 [ reached 'max' / getOption("max.print") -- omitted 2 rows ]
tb |>
  count(carrier)
# A tibble: 14 × 2
   carrier     n
   <chr>   <int>
 1 AA      26302
 2 AS        574
 3 B6      44479
 4 DL      41683
 5 EV      39819
 6 F9        473
 7 FL       1251
 8 HA        260
 9 MQ      18559
10 OO        200
11 UA      46267
12 US      16750
13 VX       4797
14 WN      11902
dt[, .(n = .N), by = carrier]
    carrier     n
     <char> <int>
 1:      AA 26302
 2:      AS   574
 3:      B6 44479
 4:      DL 41683
 5:      EV 39819
 6:      F9   473
 7:      FL  1251
 8:      HA   260
 9:      MQ 18559
10:      VX  4797
11:      WN 11902
 [ reached 'max' / getOption("max.print") -- omitted 3 rows ]
autoplot(bench_dat$counts)

Mutate

Mutate is perhaps the area where data.table feels most unlike either base R or tidyverse.

df$dist_km <- df$distance * 1.6
df[c("dist_km", "carrier", "origin", "dest")]
  dist_km carrier origin dest
1  3960.0      AA    JFK  LAX
2  3960.0      AA    JFK  LAX
3  3960.0      AA    JFK  LAX
4  1656.0      AA    LGA  PBI
5  3960.0      AA    JFK  LAX
6  3926.4      AA    EWR  LAX
 [ reached 'max' / getOption("max.print") -- omitted 253310 rows ]
tb |>
  mutate(dist_km = distance * 1.6) |>
  select(dist_km, carrier, origin, dest)
# A tibble: 253,316 × 4
   dist_km carrier origin dest 
     <dbl> <chr>   <chr>  <chr>
 1   3960  AA      JFK    LAX  
 2   3960  AA      JFK    LAX  
 3   3960  AA      JFK    LAX  
 4   1656  AA      LGA    PBI  
 5   3960  AA      JFK    LAX  
 6   3926. AA      EWR    LAX  
 7   3960  AA      JFK    LAX  
 8   3960  AA      JFK    LAX  
 9   1742. AA      JFK    MIA  
10   3875. AA      JFK    SEA  
# ℹ 253,306 more rows
# by design, dropping un-used columns:
dt[,.(dist_km = distance * 1.6, carrier, origin, dest)]
        dist_km carrier origin   dest
          <num>  <char> <char> <char>
     1:  3960.0      AA    JFK    LAX
     2:  3960.0      AA    JFK    LAX
     3:  3960.0      AA    JFK    LAX
     4:  1656.0      AA    LGA    PBI
     5:  3960.0      AA    JFK    LAX
 [ reached 'max' / getOption("max.print") -- omitted 6 rows ]
# or you can also effectively specify the columns to be retained as grouping variables
dt[,.(dist_km = distance*1.6),.(carrier,origin, dest)]
        carrier origin   dest dist_km
         <char> <char> <char>   <num>
     1:      AA    JFK    LAX  3960.0
     2:      AA    JFK    LAX  3960.0
     3:      AA    JFK    LAX  3960.0
     4:      AA    JFK    LAX  3960.0
     5:      AA    JFK    LAX  3960.0
 [ reached 'max' / getOption("max.print") -- omitted 6 rows ]
# or the := operator can modify a column in place. This allows you to change a single column in a data.table without having to re-include all the other columns. Note this permanently changes the object:

new_dt <- dt
new_dt[, dist_km := distance*1.6]
new_dt
Indices: <dest>, <month>
         year month   day dep_delay arr_delay carrier origin   dest air_time
        <int> <int> <int>     <int>     <int>  <char> <char> <char>    <int>
     1:  2014     1     1        14        13      AA    JFK    LAX      359
        distance  hour dist_km
           <int> <int>   <num>
     1:     2475     9  3960.0
 [ reached 'max' / getOption("max.print") -- omitted 10 rows ]
autoplot(bench_dat$mutates)