Filter a Microsoft.Data.Analysis.DataFrame by DateTime?

Question

Using the Microsoft.Data.Analysis package (0.18.0), I'm attempting to create a Filter query for a DateTime column and I'm receiving a System.NotSupportedException exception.

Example code:

DateTime[] birthdays = {DateTime.Parse("4/23/1990"), DateTime.Parse("4/5/1982"), DateTime.Parse("2/1/1980"), DateTime.Parse("9/10/1994") };
var birthdayColumn = new PrimitiveDataFrameColumn<DateTime>("Birthday",birthdays);
var df = new DataFrame(birthdayColumn);
var query = df["Birthday"].ElementwiseEquals(DateTime.Parse("4/5/1982"));
var theBirthday = df.Filter(query);

Exception is thrown on the 4th line:

Any ideas? How do I filter a DataFrame by a System.DateTime column?

Answer 1

I was able to work around this seeming limitation by adding a new Int64 column representing the DateTime column in Ticks, and then filtering on that column:

DateTime[] birthdays = {DateTime.Parse("4/23/1990"), DateTime.Parse("4/5/1982"), DateTime.Parse("2/1/1980"), DateTime.Parse("9/10/1994") };
var birthdayColumn = new PrimitiveDataFrameColumn<DateTime>("Birthday",birthdays);
var df = new DataFrame(birthdayColumn);
var ticks = df["Birthday"].Cast<DateTime>().Select(x => x.Ticks);
df["BirthdayTicks"] = new Int64DataFrameColumn("BirthdayTicks", ticks);
var query = df["BirthdayTicks"].ElementwiseEquals(DateTime.Parse("4/5/1982").Ticks);
var theBirthday = df.Filter(query);

Answer 2

Filter() only has overloads for Int32, Int64, and Boolean types, not DateTime. The workaround that you used is likely the best option at this time.