Logging for 'dplyr' and 'tidyr' Functions.
tidylog
The goal of tidylog is to provide feedback about dplyr and tidyr operations. It provides simple wrapper functions for almost all dplyr and tidyr functions, such as filter, mutate, select, full_join, and group_by.
Example
Load tidylog after dplyr and/or tidyr:
library("dplyr")
library("tidyr")
library("tidylog", warn.conflicts = FALSE)
Tidylog will give you feedback, for instance when filtering a data frame or adding a new variable:
filtered <- filter(mtcars, cyl == 4)
#> filter: removed 21 rows (66%), 11 rows remaining
mutated <- mutate(mtcars, new_var = wt ** 2)
#> mutate: new variable 'new_var' (double) with 29 unique values and 0% NA
Tidylog reports detailed information for joins:
joined <- left_join(nycflights13::flights, nycflights13::weather,
by = c("year", "month", "day", "origin", "hour", "time_hour"))
#> left_join: added 9 columns (temp, dewp, humid, wind_dir, wind_speed, …)
#> > rows only in nycflights13::flights 1,556
#> > rows only in nycflights13::weather ( 6,737)
#> > matched rows 335,220
#> > =========
#> > rows total 336,776
In this case, we see that 1,556 rows from the flights dataset do not have weather information.
Tidylog can be especially helpful in longer pipes:
summary <- mtcars %>%
select(mpg, cyl, hp, am) %>%
filter(mpg > 15) %>%
mutate(mpg_round = round(mpg)) %>%
group_by(cyl, mpg_round, am) %>%
tally() %>%
filter(n >= 1)
#> select: dropped 7 variables (disp, drat, wt, qsec, vs, …)
#> filter: removed 6 rows (19%), 26 rows remaining
#> mutate: new variable 'mpg_round' (double) with 15 unique values and 0% NA
#> group_by: 3 grouping variables (cyl, mpg_round, am)
#> tally: now 20 rows and 4 columns, 2 group variables remaining (cyl, mpg_round)
#> filter (grouped): no rows removed
Here, it might have been accidental that the last filter command had no effect.
Installation
Download from CRAN:
install.packages("tidylog")
Or install the development version:
devtools::install_github("elbersb/tidylog")
Benchmarks
Tidylog will add a small overhead to each function call. This can be relevant for very large datasets and especially for joins. If you want to switch off tidylog for a single long-running command, simply prefix dplyr:: or tidyr::, such as in dplyr::left_join. See this vignette for more information.
More examples
filter, distinct, drop_na
a <- filter(mtcars, mpg > 20)
#> filter: removed 18 rows (56%), 14 rows remaining
b <- filter(mtcars, mpg > 100)
#> filter: removed all rows (100%)
c <- filter(mtcars, mpg > 0)
#> filter: no rows removed
d <- filter_at(mtcars, vars(starts_with("d")), any_vars((. %% 2) == 0))
#> filter_at: removed 19 rows (59%), 13 rows remaining
e <- distinct(mtcars)
#> distinct: no rows removed
f <- distinct_at(mtcars, vars(vs:carb))
#> distinct_at: removed 18 rows (56%), 14 rows remaining
g <- top_n(mtcars, 2, am)
#> top_n: removed 19 rows (59%), 13 rows remaining
i <- sample_frac(mtcars, 0.5)
#> sample_frac: removed 16 rows (50%), 16 rows remaining
j <- drop_na(airquality)
#> drop_na: removed 42 rows (27%), 111 rows remaining
k <- drop_na(airquality, Ozone)
#> drop_na: removed 37 rows (24%), 116 rows remaining
mutate, transmute, replace_na, fill
a <- mutate(mtcars, new_var = 1)
#> mutate: new variable 'new_var' (double) with one unique value and 0% NA
b <- mutate(mtcars, new_var = runif(n()))
#> mutate: new variable 'new_var' (double) with 32 unique values and 0% NA
c <- mutate(mtcars, new_var = NA)
#> mutate: new variable 'new_var' (logical) with one unique value and 100% NA
d <- mutate_at(mtcars, vars(mpg, gear, drat), round)
#> mutate_at: changed 28 values (88%) of 'mpg' (0 new NAs)
#> changed 31 values (97%) of 'drat' (0 new NAs)
e <- mutate(mtcars, am_factor = as.factor(am))
#> mutate: new variable 'am_factor' (factor) with 2 unique values and 0% NA
f <- mutate(mtcars, am = as.ordered(am))
#> mutate: converted 'am' from double to ordered factor (0 new NA)
g <- mutate(mtcars, am = ifelse(am == 1, NA, am))
#> mutate: changed 13 values (41%) of 'am' (13 new NAs)
h <- mutate(mtcars, am = recode(am, `0` = "zero", `1` = NA_character_))
#> mutate: converted 'am' from double to character (13 new NA)
i <- transmute(mtcars, mpg = mpg * 2, gear = gear + 1, new_var = vs + am)
#> transmute: dropped 9 variables (cyl, disp, hp, drat, wt, …)
#> transmute: dropped 9 variables (cyl, disp, hp, drat, wt, …)
#> changed 32 values (100%) of 'mpg' (0 new NAs)
#> changed 32 values (100%) of 'gear' (0 new NAs)
#> new variable 'new_var' (double) with 3 unique values and 0% NA
j <- replace_na(airquality, list(Solar.R = 1))
#> replace_na: changed 7 values (5%) of 'Solar.R' (7 fewer NAs)
k <- fill(airquality, Ozone)
#> fill: changed 37 values (24%) of 'Ozone' (37 fewer NAs)
joins
For joins, tidylog provides more detailed information. For any join, tidylog will show the number of rows that are only present in x (the first dataframe), only present in y (the second dataframe), and rows that have been matched. Numbers in parentheses indicate that these rows are not included in the result. Tidylog will also indicate whether any rows were duplicated (which is often unintentional):
x <- tibble(a = 1:2)
y <- tibble(a = c(1, 1, 2), b = 1:3) # 1 is duplicated
j <- left_join(x, y, by = "a")
#> left_join: added one column (b)
#> > rows only in x 0
#> > rows only in y (0)
#> > matched rows 3 (includes duplicates)
#> > ===
#> > rows total 3
More examples:
a <- left_join(band_members, band_instruments, by = "name")
#> left_join: added one column (plays)
#> > rows only in band_members 1
#> > rows only in band_instruments (1)
#> > matched rows 2
#> > ===
#> > rows total 3
b <- full_join(band_members, band_instruments, by = "name")
#> full_join: added one column (plays)
#> > rows only in band_members 1
#> > rows only in band_instruments 1
#> > matched rows 2
#> > ===
#> > rows total 4
c <- anti_join(band_members, band_instruments, by = "name")
#> anti_join: added no columns
#> > rows only in band_members 1
#> > rows only in band_instruments (1)
#> > matched rows (2)
#> > ===
#> > rows total 1
Because tidylog needs to perform two additional joins behind the scenes to report this information, the overhead will be larger than for the other tidylog functions (especially with large datasets).
select, relocate, rename
a <- select(mtcars, mpg, wt)
#> select: dropped 9 variables (cyl, disp, hp, drat, qsec, …)
b <- select_if(mtcars, is.character)
#> select_if: dropped all variables
c <- relocate(mtcars, hp)
#> relocate: columns reordered (hp, mpg, cyl, disp, drat, …)
d <- select(mtcars, a = wt, b = mpg)
#> select: renamed 2 variables (a, b) and dropped 9 variables
e <- rename(mtcars, miles_per_gallon = mpg)
#> rename: renamed one variable (miles_per_gallon)
f <- rename_with(mtcars, toupper)
#> rename_with: renamed 11 variables (MPG, CYL, DISP, HP, DRAT, …)
summarize
a <- mtcars %>%
group_by(cyl, carb) %>%
summarize(total_weight = sum(wt))
#> group_by: 2 grouping variables (cyl, carb)
#> summarize: now 9 rows and 3 columns, one group variable remaining (cyl)
b <- iris %>%
group_by(Species) %>%
summarize_all(list(min, max))
#> group_by: one grouping variable (Species)
#> summarize_all: now 3 rows and 9 columns, ungrouped
tally, count, add_tally, add_count
a <- mtcars %>% group_by(gear, carb) %>% tally
#> group_by: 2 grouping variables (gear, carb)
#> tally: now 11 rows and 3 columns, one group variable remaining (gear)
b <- mtcars %>% group_by(gear, carb) %>% add_tally()
#> group_by: 2 grouping variables (gear, carb)
#> add_tally (grouped): new variable 'n' (integer) with 5 unique values and 0% NA
c <- mtcars %>% count(gear, carb)
#> count: now 11 rows and 3 columns, ungrouped
d <- mtcars %>% add_count(gear, carb, name = "count")
#> add_count: new variable 'count' (integer) with 5 unique values and 0% NA
pivot_longer, pivot_wider
longer <- mtcars %>%
mutate(id = 1:n()) %>%
pivot_longer(-id, names_to = "var", values_to = "value")
#> mutate: new variable 'id' (integer) with 32 unique values and 0% NA
#> pivot_longer: reorganized (mpg, cyl, disp, hp, drat, …) into (var, value) [was 32x12, now 352x3]
wider <- longer %>%
pivot_wider(names_from = var, values_from = value)
#> pivot_wider: reorganized (var, value) into (mpg, cyl, disp, hp, drat, …) [was
#> 352x3, now 32x12]
Tidylog also supports gather and spread.
Turning logging off, registering additional loggers
To turn off the output for just a particular function call, you can simply call the dplyr and tidyr functions directly, e.g. dplyr::filter or tidyr::drop_na.
To turn off the output more permanently, set the global option tidylog.display to an empty list:
options("tidylog.display" = list()) # turn off
a <- filter(mtcars, mpg > 20)
options("tidylog.display" = NULL) # turn on
a <- filter(mtcars, mpg > 20)
#> filter: removed 18 rows (56%), 14 rows remaining
This option can also be used to register additional loggers. The option tidylog.display expects a list of functions. By default (when tidylog.display is set to NULL), tidylog will use the message function to display the output, but if you prefer a more colorful output, simply overwrite the option:
library("crayon") # for terminal colors
crayon <- function(x) cat(red$bold(x), sep = "\n")
options("tidylog.display" = list(crayon))
a <- filter(mtcars, mpg > 20)
#> filter: removed 18 rows (56%), 14 rows remaining
To print the output both to the screen and to a file, you could use:
log_to_file <- function(text) cat(text, file = "log.txt", sep = "\n", append = TRUE)
options("tidylog.display" = list(message, log_to_file))
a <- filter(mtcars, mpg > 20)
#> filter: removed 18 rows (56%), 14 rows remaining
Namespace conflicts
Tidylog redefines several of the functions exported by dplyr and tidyr, so it should be loaded last, otherwise there will be no output. A more explicit way to resolve namespace conflicts is to use the conflicted package:
library("dplyr")
library("tidyr")
library("tidylog")
library("conflicted")
for (f in getNamespaceExports("tidylog")) {
conflicted::conflict_prefer(f, "tidylog", quiet = TRUE)
}