Update: gather() is no longer maintained and it is recommended that pivot_longer() is used. For details, see here.
The gather() function is somewhat complementary to what the spread() function does.
gather(data, key, value, -variable)
where - stands for “exclude”, meaning that variable is excluded in this case.
Working example:
stocks <- tibble(
time = as.Date('2009-01-01') + 0:9,
X = rnorm(10, 0, 1),
Y = rnorm(10, 0, 2),
Z = rnorm(10, 0, 4)
)
stocks
## # A tibble: 10 × 4
## time X Y Z
## <date> <dbl> <dbl> <dbl>
## 1 2009-01-01 0.208 1.14 -1.67
## 2 2009-01-02 2.40 1.74 -1.14
## 3 2009-01-03 -0.424 2.01 -1.03
## 4 2009-01-04 0.0343 -0.978 3.50
## 5 2009-01-05 0.814 0.280 -2.14
## 6 2009-01-06 -0.361 -0.404 -1.57
## 7 2009-01-07 -1.59 0.649 1.90
## 8 2009-01-08 -1.15 3.25 4.10
## 9 2009-01-09 -1.20 -0.818 4.01
## 10 2009-01-10 2.23 2.25 0.777
In this table, to put the variables (X, Y, and Z) into a new variable “stock”, and al the values to a new variable “price”, it can be achieved via below:
gather(stocks, key = "stock", value = "price", -time)
## # A tibble: 30 × 3
## time stock price
## <date> <chr> <dbl>
## 1 2009-01-01 X 0.208
## 2 2009-01-02 X 2.40
## 3 2009-01-03 X -0.424
## 4 2009-01-04 X 0.0343
## 5 2009-01-05 X 0.814
## 6 2009-01-06 X -0.361
## 7 2009-01-07 X -1.59
## 8 2009-01-08 X -1.15
## 9 2009-01-09 X -1.20
## 10 2009-01-10 X 2.23
## # … with 20 more rows