# given this structure:
array_of_hashes = [
{:callid => '1', :foo => '1', :created_at => '2007-09-21 11:45:15'},
{:callid => '1', :foo => '2', :created_at => '2007-09-21 11:45:16'},
{:callid => '2', :foo => '1', :created_at => '2007-09-21 11:45:19'},
{:callid => '2', :foo => '2', :created_at => '2007-09-21 11:45:20'},
{:callid => '3', :foo => '1', :created_at => '2007-09-21 11:45:21'}
]
# group using this:
array_of_arrays_of_hashes = returning([]) do |grouped_array|
array_of_hashes.map {|x| x[:callid]}.uniq.each do |callid|
grouped_array[callid.to_i] = array_of_hashes.select {|y| y[:callid] == callid}
end
end.compact
# and this is the result:
array_of_arrays_of_hashes == [
[{:callid=>"1", :foo=>"1", :created_at=>"2007-09-21 11:45:15"},
{:callid=>"1", :foo=>"2", :created_at=>"2007-09-21 11:45:16"}],
[{:callid=>"2", :foo=>"1", :created_at=>"2007-09-21 11:45:19"},
{:callid=>"2", :foo=>"2", :created_at=>"2007-09-21 11:45:20"}],
[{:callid=>"3", :foo=>"1", :created_at=>"2007-09-21 11:45:21"}]
]
Refactorings
No refactoring yet !
dudemeister
October 1, 2007, October 01, 2007 15:30, permalink
No rating.
Login to rate!
With group_by (found e.g. in rails), and methodphitamine:
# ... # group using this: array_of_arrays_of_hashes = array_of_hashes.group_by(&its[:callid]).values
hungryblank
October 1, 2007, October 01, 2007 15:40, permalink
No rating.
Login to rate!
I thought for you it was important array_of_arrays_of_hashes[idx] would return the array with the idx caller id.
But is not because after you apply the compact method and this breaks the relationship between array_of_arrays_of_hashes index and callid.
In this case the code below should be enough.
array_of_arrays_of_hashes = array_of_hashes.map { |e| e[:callid] }.uniq.inject([]) do |grouped_array, callid|
grouped_array << array_of_hashes.select { |h| h[:callid] == callid }
end
This chunk of code takes an array of hashes and groups it by "callid". Does this by creating a temporary array of unique callid's and then feeding them into a new "grouped" array using the callid as an index. The result is an array of arrays of hashes as per the example after the grouping code. This works, but I have to think that there is a cleaner method of restructuring this data in this way.