2 Ways to Return Multiple Values with sapply in R

readme.txt

2 Ways to Return Multiple Values with sapply
---

Demo at http://www.r-fiddle.org/#/fiddle?id=4pR9ndzC&version=2

Assume we have the following data:

        Date Value
1 2015-05-01    73
2 2015-05-01    94
3 2015-05-02   100
4 2015-05-02    83
5 2015-05-03    47
6 2015-05-04    24

We want to convert this data by day, into the following format, for easy display in a chart:

Count Total       Date
    2   167 2015-05-01
    2   183 2015-05-02
    1    47 2015-05-03
    2   100 2015-05-04
    2   123 2015-05-05
    1    32 2015-05-06

We can use sapply over the data.frame to convert the data. During each loop iteration in saaply, we can either populate an outside data.frame with a new row or we can return a vector of the result values in each iteration. See Method 1 and Method 2 in the code example below.

With our data formatted, we can plot a chart as shown in the example below.

sapply.R

# Create some sample data.
dates <- seq(as.Date("2015/5/1"), as.Date("2015/6/30"), length.out=100)
values <- sample(1:100, length(dates))

# Create a data.frame of data.
data <- data.frame(Date = dates, Value = values)

# Split by day.
dataByDay <- split(data, data$Date)

# Get list of data counts and sum per day.
# Method 1.
dataCountsByDay1 <- data.frame()
x <- sapply(dataByDay, function(r) {
  df <- data.frame(r)
  row <- data.frame(Count = nrow(df), Total = sum(df$Value), Date = df[1,]$Date)
  dataCountsByDay1 <<- rbind(dataCountsByDay1, row)
})

# Get list of data counts and sum per day.
# Method 2.
dataCountsByDay2 <- as.data.frame(t(sapply(dataByDay, function(r) {
  df <- data.frame(r)
  c(Count = nrow(df), Total = sum(df$Value))
})))
dataCountsByDay2$Date <- as.Date(row.names(dataCountsByDay2))

# Verify they're both equivalent.
dataCountsByDay1$Count == dataCountsByDay2$Count
dataCountsByDay1$Total == dataCountsByDay2$Total
as.character(dataCountsByDay1$Date) == as.character(dataCountsByDay2$Date)
nrow(dataCountsByDay1) == nrow(dataCountsByDay2)

values.png