Joins

Imagine we have two data frames x and y:

  x1 x2
1  A  1
2  B  2
3  C  3

  x1    x3
1  A  TRUE
2  B FALSE
3  D  TRUE

We want to join them along the categorical variable x1and end up with a new data frame that has all three variables:

inner_join(x, y, by = "x1")

  x1 x2    x3
1  A  1  TRUE
2  B  2 FALSE

full_join(x, y, by = "x1")

  x1 x2    x3
1  A  1  TRUE
2  B  2 FALSE
3  C  3    NA
4  D NA  TRUE

• dplyr's operations to join data sets are inspired by SQL (Structured Query Language), which is used to query large databases.
• If a value is missing during the join, NA (R code for missing value) is inserted.

In our example, both data frames x and y the had the “join by” variable named the same: x1. Hence:

inner_join(x, y, by = "x1")

If the “join by” variable has different names in the data frames. Ex:

• name a in data frame x
• name b in data frame y

then run: inner_join(x, y, by = c("a"="b")).

Note it is

• by = c("a"="b") and not
• by = c("a"=="b")

See by in the “Argument” section of the help file ?join.

• See ?join for a list of the different types of joins
• This illustration succinctly summarizes all of them.

At the beginning of Lec03.R, I load all necessary packages at the beginning.

This is good practice people know which packages they need installed. For example, I didn't know I needed a particular package before I boarded a 4 hour flight because the package loading was buried in the file.