How to average similar tuples in an array of tuples in Swift - arrays

Question

I really need your help. I have an array of tuples that looks like this:
[("07-21-2016", 5), ("07-21-2016", 1), ("07-21-2016", 2), ("07-21-2016", 3), ("07-21-2016", 4), ("07-21-2016", 5), ("07-20-2016", 6), ("07-20-2016", 5), ("07-19-2016", 5)]
I need to take all the tuples with the same date and average them out. So at the end it would look like:
[("07-21-2016", 33.3), ("07-20-2016", 5.5), ("07-19-2016", 5)]
Does anyone know how to do this?

Answer 1

let array = [("07-21-2016", 5), ("07-21-2016", 1), ("07-21-2016", 2), ("07-21-2016", 3), ("07-21-2016", 4), ("07-21-2016", 5), ("07-20-2016", 6), ("07-20-2016", 5), ("07-19-2016", 5)]
// Create dictionary to hold mapping of date to array of values
var dict = [String: [Double]]()
// use forEach to add each value to the array for each key
array.forEach {(date, num) in dict[date] = (dict[date] ?? []) + [Double(num)]}
// use map with reduce to find the average of each value and return a tuple
// containing the date and the average value
let result = dict.map {(date, nums) in (date, nums.reduce(0, combine: +) / Double(nums.count))}
print(result)
Output:
[("07-20-2016", 5.5), ("07-19-2016", 5.0), ("07-21-2016", 3.3333333333333335)]
Explanation:
array.forEach {(date, num) in dict[date] = (dict[date] ?? []) + [Double(num)]}
forEach takes each tuple of the array, looks up the array of values corresponding to dict[date] and appends the new num to that array. If dict[date] returns nil, then this is the first time we've seen this key, so use the nil coalescing operator ?? to return an empty array [] and append the new value to that.
At the end of this, the contents of dict is:
["07-20-2016": [6.0, 5.0], "07-19-2016": [5.0], "07-21-2016": [5.0, 1.0, 2.0, 3.0, 4.0, 5.0]]
let result = dict.map {(date, nums) in (date, nums.reduce(0, combine: +) / Double(nums.count))}
When map is applied to a dictionary, it takes each (key, value) pair and creates a new value based upon that. The end result of map is a new array of the values it returns. In this case, the value returned for each iteration of map is a tuple containing the date and the average of the numbers associated with that date.
nums.reduce(0, combine: +)
This sums the values in the nums array. reduce takes an initial value (0 in this case) and a closure that will be evaluated for each value in the nums array. Each iteration of reduce takes the current running total and the next value in nums and sums them. This sum is then divided by Double(nums.count) to produce the average. Finally, map returns (date, avg) which produces the final result.

Answer 2

Here's one way (using a dictionary to collate the numbers):
let dateValues = [("07-21-2016", 5), ("07-21-2016", 1), ("07-21-2016", 2), ("07-21-2016", 3), ("07-21-2016", 4), ("07-21-2016", 5), ("07-20-2016", 6), ("07-20-2016", 5), ("07-19-2016", 5)]
var averages:[String:(Int,Int)] = [:]
for (date,value) in dateValues
{
averages[date] = averages[date] ?? (0,0)
averages[date] = (averages[date]!.0 + value, averages[date]!.1 + 1)
}
let averagePerDate = averages.map{($0,Float($1.0)/Float($1.1))}.sort{$0.0>$1.0}
print(averagePerDate)
// [("07-21-2016", 3.33333325), ("07-20-2016", 5.5), ("07-19-2016", 5.0)]
And a more concise one using sets:
let dateList = dateValues.reduce( Set<String>(), combine: { $0.union(Set([$1.0])) })
let dateData = dateList.map{ date in return (date, dateValues.filter({$0.0==date}).map{$0.1}) }
let dateCounts = dateData.map{ ($0, $1.reduce(0,combine:+), Float($1.count) ) }
let dateAverages = dateCounts.map{ ($0, Float($1/$2) ) }.sort{$0.0>$1.0}
print(dateAverages)