When it comes to rbind:.
It only works when the column names in the binding datasets are the same.
Section 4.4 Merging Data
What if we want our two datasets together?
Subsection 4.4.1 Binding our data
We have data for the first 20 games for the mets and yankees but they’re split into the first 10 and second 10 games. What if we want to see all 20 games for each individual team? We can use a command in base R called
rbind. The r stands for “rows” so this command is literally telling R to “bind the rows.”
When combining datasets, no matter which method, it is imperative that we check our data before and after to make sure the combining process worked how we anticipated. In this case, each of our original datasets contain 10 rows, which means our combined datasets should have 20 rows. We will be checking using
nrow, and any number besides 20 means that there is an issue.
# Mets
nrow(mets) # Seeing how many rows are in mets
nrow(mets_second_ten) # Seeing how many rows are in mets_second_ten
mets_all <- rbind(mets, mets_second_ten) # Combining the rows of mets and mets_second_ten
nrow(mets_all) # Seeing how many rows are in mets_all
# Yankees
nrow(yankees)
nrow(yankees_second_ten)
yankees_all <- rbind(yankees, yankees_second_ten)
[1] 10 [1] 10 [1] 20 [1] 10 [1] 10
Success! Both of our datasets have 20 rows.
Subsection 4.4.2 Joining Data
Our two datasets, mets_all and yankees_all now have the data from the first 20 games of the 2025 MLB season, since we successfully combined our data. Now, what if we want one dataset that had the scores of each of the first 20 games for the Mets and Yankees?
There are a few different ways to do this. The first is to use the
merge command which is part of base R. To check that our data has merged successfully, we need to make sure there are 20 rows and 3 columns: Game, Mets_Score, and Yankees_Score.
baseball_data <- merge(yankees_all, mets_all, by = "Game")
baseball_data
Game Yankees_Score Mets_Score 1 1 3 4 2 2 8 20 3 3 2 12 4 4 3 5 5 5 2 3 6 6 6 9 7 7 5 9 8 8 3 10 9 9 2 4 10 10 7 2 11 11 0 10 12 12 4 0 13 13 1 7 14 14 8 1 15 15 4 8 16 16 4 5 17 17 4 3 18 18 4 3 19 19 6 4 20 20 1 1
Fantastic! We now have one complete dataset with the scores of the Mets and Yankees in the first 20 games.
Subsubsection 4.4.2.1 Using SQL Joins
We previously used the
merge command, however, R also allows for SQL code! Here is a quick overview of the main joins using SQL in R:
-
inner_join: Keeps only rows that match in both datasets
-
Overlap
-
-
full_join: Keeps all rows from both datasets, matching when possible
-
Everything
-
-
left_join: Keeps all rows from the left dataset, and matches from the right when possible
-
All left
-
-
right_join: Keeps all rows from the right dataset, and matches from the left when possible
-
All right
-
-
anti_join: Keeps rows from the left dataset that do NOT appear in the right
-
What’s missing
-
In this case, we are going to use
inner_join, since we know that both of our datasets have the same 20 games.
library(tidyverse)
baseball_data_sql <- inner_join(yankees_all, mets_all, by = "Game")
nrow(baseball_data_sql) # Checks the number of rows
ncol(baseball_data_sql) # Checks the number of columns
baseball_data_sql
[1] 20 [1] 3 Game Yankees_Score Mets_Score 1 1 3 4 2 2 8 20 3 3 2 12 4 4 3 5 5 5 2 3 6 6 6 9 7 7 5 9 8 8 3 10 9 9 2 4 10 10 7 2 11 11 0 10 12 12 4 0 13 13 1 7 14 14 8 1 15 15 4 8 16 16 4 5 17 17 4 3 18 18 4 3 19 19 6 4 20 20 1 1
One of the beautiful things in R is that you can get the same results using different code - just like we did here. Our two baseball datasets are identical even though we used two different commands to get them. This is an example of how R puts the r in Artist.
Subsection 4.4.3 Wide Format
Right now, our data is considered wide format, that is, there are 20 rows, and separate columns for the scores of the Yankees and scores of the Mets. In some cases, we may need our data to be in a long format. In this case, there will be 40 rows instead of 20, since each Yankees score and each Mets score will represent a row, instead of each game representing a row.
Using the
pivot_longer function from the tidyverse package, we can turn our data from wide to long.
# Lets say we want to turn out data from wide format into long format, so we can run
baseball_data_long <- baseball_data |> pivot_longer(cols = c(Yankees_Score, Mets_Score),
names_to = "Team",
values_to = "Score")
nrow(baseball_data_long)
baseball_data_long
[1] 40
# A tibble: 40 × 3
Game Team Score
<int> <chr> <dbl>
1 1 Yankees_Score 3
2 1 Mets_Score 4
3 2 Yankees_Score 8
4 2 Mets_Score 20
5 3 Yankees_Score 2
6 3 Mets_Score 12
7 4 Yankees_Score 3
8 4 Mets_Score 5
9 5 Yankees_Score 2
10 5 Mets_Score 3
# ℹ 30 more rows
Subsection 4.4.4 Long Format (Reverse Demo)
Our original baseball_data (and baseball_data_sql) are in wide format already so we don’t have to do this, but just in case you ever start with long data and want to convert it to wide, here is some code to help. Below, we are turning our baseball_data_long into a wide format.
# Lets say we want to turn out data from wide format into long format, so we can run
baseball_data_wide <- baseball_data_long |> pivot_wider(names_from = "Team",
values_from = "Score")
nrow(baseball_data_wide)
baseball_data_wide
[1] 20
# A tibble: 20 × 3
Game Yankees_Score Mets_Score
<int> <dbl> <dbl>
1 1 3 4
2 2 8 20
3 3 2 12
4 4 3 5
5 5 2 3
6 6 6 9
7 7 5 9
8 8 3 10
9 9 2 4
10 10 7 2
11 11 0 10
12 12 4 0
13 13 1 7
14 14 8 1
15 15 4 8
16 16 4 5
17 17 4 3
18 18 4 3
19 19 6 4
20 20 1 1
