lesson07

suppressPackageStartupMessages(library(tidyverse))

Define three simple data frames. Also define names2 and plays2 as shown in DataCamp.

t1 = data_frame(alpha = letters[1:6], num = 1:6) 
t2 = data_frame(alpha = letters[4:10], num = 4:10)

joins

multi column join

joining names2 and plays2

left_join(names2,plays2) # still works, by is defined automatically
left_join(names2,plays2, by=c("name","surname"))

inner join

now, back to t1 and t2

t1
t2
left_join(t1,t2, by="alpha")
inner_join(t1, t2, by = "alpha")

left-join

left_join(t1, t2, by = "alpha")

right-join

you do the join

full-join

if not set, the suffix is assigned as: x for LHS, y for RHS

full_join(t1, t2, by = "alpha")
full_join(t1, t2, by = "alpha", suffix = c("_t1", "_t2"))

about is.na()

full_join(t1, t2, by = "alpha", suffix = c("_t1", "_t2")) %>%
  mutate(missing=is.na(num_t1))

na.rm is an option of sum() or mean() functions (actually many functions rely on it) to remove NA. But, is.na() is a standalone function which returns a logical vector.

birvector <- c(4,NA,6,8)
is.na(birvector)
sum(birvector)
sum(birvector, na.rm=TRUE)

pipe on right

By the way, if needed be, the table coming from pipe can be used on right.

t1 %>%
  left_join(t2, . , by="alpha")

intersect

uses whole row

t1
t2
t3

semi_join(t1,t3, by="alpha")
intersect(t1,t3)

data.frame vs data_frame

compare t1 and t1old

t1old = data.frame("alpha-beta" = letters[1:6], num = 1:6)
t1new = data_frame("alpha-beta" = letters[1:6], num = 1:6)
t1 = data_frame(alpha = letters[1:6], num = 1:6) 
t2 = data_frame(alpha = letters[4:10], num = 4:10)

t1old
t1

you find an example of recycling of vectors

bind

bind_rows(tablo1=t1,tablo2=t2,.id="table")

flights tables

nycflights13 tables