Returning the Most Frequent Value between Date & Time to Aggregate a Count

[Anonymous]

Our unscheduled appointments do not have a location and return a value of NA. I would like to graph total filled appointments / total available appointments by location. This creates a challenge as a provider can change locations around Noon and I need a way around NA. I would like to have a way to compare the various locations for a timeperiod of 8am to Noon and return back the most frequent value that isn't NA.

Any ideas would be appreciated! Below is an example of that data 

Appointment	Provider	Location	Requested Result
7/13/2022 8:00	Doctor	Location1	Location1
7/13/2022 8:20	Doctor	NA	Location1
7/13/2022 13:00	Doctor	Location2	Location2
7/13/2022 13:20	Doctor	NA	Location2
7/13/2022 13:20	Doctor	Location2	Location2
7/13/2022 16:20	Doctor	NA	Location2

[jgeddes · Accepted solution]

If I understand correctly now, you want to replace the NA values with the most common location for the date and whether it is before or after noon.

I created the following dataset...

and then did the following in Power Query

let
Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WMtc3NNY3MjAyUrCwMjBQ0lFyyU8uyS8CMnzykxNLMvPzDFHYsTqoeoxQ9Pg54lFsaIzDBiMUNoYmPFYQVE2UFWYkWWFuZQxS7VdaVJyKpM4YhY2pxxSLHpCrlGJjAQ==", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Appointment = _t, Provider = _t, Location = _t, #"Requested Result" = _t]),
#"Changed Type" = Table.TransformColumnTypes(Source,{{"Appointment", type datetime}, {"Provider", type text}, {"Location", type text}, {"Requested Result", type text}}),
#"Inserted Date" = Table.AddColumn(#"Changed Type", "Date", each DateTime.Date([Appointment]), type date),
#"Added Custom" = Table.AddColumn(#"Inserted Date", "isBeforeNoon", each if Time.From([Appointment]) <= #time(12,0,0) then "Before" else "After"),
#"Added Custom1" = Table.AddColumn(#"Added Custom", "Count of Location", each List.Count(Table.SelectRows(#"Added Custom", (x) => x[Location] <> "NA" and x[Date] = [Date] and x[isBeforeNoon] = [isBeforeNoon] and x[Location] = [Location])[Location])),
#"Added Custom2" = Table.AddColumn(#"Added Custom1", "Max Count", each List.Max(Table.SelectRows(#"Added Custom1", (x) => x[Date] = [Date] and x[isBeforeNoon] = [isBeforeNoon])[Count of Location])),
#"Added Custom3" = Table.AddColumn(#"Added Custom2", "Most Common Location", each List.Max(Table.SelectRows(#"Added Custom2", (x) => x[Date] = [Date] and x[isBeforeNoon] = [isBeforeNoon] and x[Count of Location] = [Max Count])[Location])),
#"Added Custom4" = Table.AddColumn(#"Added Custom3", "adjustedLocation", each if [Location] = "NA" then [Most Common Location] else [Location]),
#"Removed Columns" = Table.RemoveColumns(#"Added Custom4",{"Requested Result", "Count of Location", "Max Count", "Most Common Location"})
in
#"Removed Columns"

 to end up with

There are likely other ways to get this result but this should get you started.

[Anonymous]

This is exactly what I was looking for. Thank you so much!

[jgeddes]

I would start by creating a column that tests if the appointment time is before or after noon. 

Once that column is in place you can use it in CALCULATE measures.

isBeforeNoon = 
var _time =
TIMEVALUE('Table (3)'[Appointment])
return
IF(
    _time < TIME(12,0,0),
    "Before",
    "After"
)

[Anonymous]

That is very helpful, I am stuck a step behind that on the fact that my data has NAs for locations. I was wondering if there is some sort of logic to assign a location where a NA exists in my direct query. My only thought to pick a location would be the most frequent in a time period