I was hoping this could be done directly upon index-time, but I suppose I could save this as a file and then re-ingest it. Thank you for this marvellous piece of work! I aspire to be able to manipulate data as you have done. Appreciated ... View more

My problem is that the actual year and the month are shown within every row of the 1st column, but then actual days are shown as column headers. Within the intersection of each ("day") column header with each row(month/year) is a value. The field name of that value is not shown in the column header(which shows the day). The field is implied to be a count statistic representing the the number of occurrences for an event occurring on that day. These are occurrences are then categorised by the row values of the 2nd column. Essentially, this means that every single cell, represents a distinct event. How do you parse that? To make things more complicated, the City and Totals column would not be a distinct event, but would be common to all events on the same row. Using the input example in my original question, my goal would be to obtain event data like this: time::2020/01/01airport::New Yorktotal::5impliedcount::1 _time::2020/01/01airport::PARIS____total::10impliedcount::0 time::2020/01/01airport::LONDONtotal::15impliedcount::6 time::2020/01/02airport::NEW YORKtotal::5_impliedcount::0 _time::2020/01/02airport::PARIS_____total::10impliedcount::5 _time::2020/01/02airport::LONDONtotal::15impliedcount::4 _time::2020/01/03airport::NEW YORKtotal::5impliedcount::0 _time::2020/01/03airport::PARIS_____total::10impliedcount::5 _time::2020/01/03airport::LONDONtotal::15___impliedcount::6 etc... ... View more