|
| 1 | +--- |
| 2 | +title: Calendar-Based Time Intelligence Is Out Of This World |
| 3 | +description: Darian Calendar using Calendar-Based Time Intelligence |
| 4 | +image: /assets/images/blog/2025/2025-11-20-CalendarTimeIntel/hero.jpg |
| 5 | +date: |
| 6 | + created: 2025-11-20 |
| 7 | +authors: |
| 8 | + - jDuddy |
| 9 | +comments: true |
| 10 | +categories: |
| 11 | + - Data Modelling |
| 12 | +slug: posts/CalendarTimeIntel |
| 13 | +links: |
| 14 | + - MS Docs: https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-time-intelligence#calendar-based-time-intelligence-preview |
| 15 | + - SQLBI: https://www.sqlbi.com/articles/introducing-calendar-based-time-intelligence-in-dax |
| 16 | +--- |
| 17 | + |
| 18 | +Microsoft recently released [Calendar-based time intelligence](https://learn.microsoft.com/en-us/power-bi/transform-model/desktop-time-intelligence#calendar-based-time-intelligence-preview) and it is out of this world. So much so lets try and make a Martian Calendar. |
| 19 | + |
| 20 | +## Calendar-Based Time Intelligence |
| 21 | + |
| 22 | +Calendar-Based Time Intelligence allows you to add metadata to your Date Dimension, specifying fields as periods (Year, Month, Quarter, **Week**), and you can define one of more different calendars for the same dimension. But why is this helpful? |
| 23 | + |
| 24 | +- **Any Calendar**: Not limited to single Gregorian, you can have Shifted Gregorian, 445, or something more exotic |
| 25 | +- **Sparse dates**: Classic time intelligence requires complete date columns with no missing dates between first and last dates. Calendar-based time intelligence operates on dates as-is, so if your stores are closed on weekends, you can skip those days entirely |
| 26 | +- **Week-based calculations**: Direct support for week granularity functions like `TOTALWTD()` |
| 27 | +- **Optimized queries**: If you filter by a defined period (Year, Month etc.) the engine can filter by the period (`Year: {2021, 2022, 2023}`), rather than falling back to date (`Date: {01-01-2021,....,31-12-2023}`), resulting in smaller cache and faster queries |
| 28 | + |
| 29 | +## Martian Calendar |
| 30 | + |
| 31 | +On Mars a day (Sol) is 24 hours, 39 minutes, and 35 seconds, nearly 40 minutes longer than a day on Earth, and the Martian year is 668.5907 sols (686.9711 Earth days). One proposed calendar for Mars is the [Darian calendar](https://en.wikipedia.org/wiki/Darian_calendar), which splits the year into 24 months with varying lengths. |
| 32 | + |
| 33 | +=== "Darian Calendar" |
| 34 | + |
| 35 | + |MartianYear | MartianMonth | SolOfMonth | MonthName | Date | SolsSinceEpoch | |
| 36 | + |---|---|---|---|---|---| |
| 37 | + |200 | 1 | 1 | Sagittarius | 1-Sagittarius-200 | 1 | |
| 38 | + |200 | 1 | 2 | Sagittarius | 2-Sagittarius-200 | 2 | |
| 39 | + |200 | 1 | 3 | Sagittarius | 3-Sagittarius-200 | 3 | |
| 40 | + |200 | 1 | 4 | Sagittarius | 4-Sagittarius-200 | 4 | |
| 41 | + |200 | 1 | 5 | Sagittarius | 5-Sagittarius-200 | 5 | |
| 42 | + |
| 43 | +=== "DAX" |
| 44 | + |
| 45 | + ```dax |
| 46 | + Dates = |
| 47 | + VAR StartMartianYear = 200 // Adjust as needed |
| 48 | + VAR EndMartianYear = 220 |
| 49 | + VAR EarthEpoch = DATE(1609, 3, 11) // Telescopic Epoch (Start of Mars Year 0) |
| 50 | + VAR SolsPerEarthDay = 1.02749125 |
| 51 | + VAR StandardYearSols = 668 // Sols in a standard Martian Year (Darian Calendar) |
| 52 | + VAR LeapYearSols = 669 // Sols in a leap Martian Year (Darian Calendar) |
| 53 | + |
| 54 | + VAR MonthNames = |
| 55 | + DATATABLE( |
| 56 | + "MartianMonth", INTEGER, |
| 57 | + "MonthName", STRING, |
| 58 | + { |
| 59 | + {1, "Sagittarius"}, {2, "Dhanus"}, {3, "Capricornus"}, {4, "Makara"}, |
| 60 | + {5, "Aquarius"}, {6, "Kumbha"}, {7, "Pisces"}, {8, "Mina"}, |
| 61 | + {9, "Aries"}, {10, "Mesha"}, {11, "Taurus"}, {12, "Rishabha"}, |
| 62 | + {13, "Gemini"}, {14, "Mithuna"}, {15, "Cancer"}, {16, "Karka"}, |
| 63 | + {17, "Leo"}, {18, "Simha"}, {19, "Virgo"}, {20, "Kanya"}, |
| 64 | + {21, "Libra"}, {22, "Tula"}, {23, "Scorpius"}, {24, "Vrishika"} |
| 65 | + } |
| 66 | + ) |
| 67 | + |
| 68 | + VAR IsLeapYear = |
| 69 | + SELECTCOLUMNS( |
| 70 | + GENERATESERIES(StartMartianYear, EndMartianYear, 1), |
| 71 | + "MartianYear", [Value], |
| 72 | + // Your Leap Rule: Odd years OR years ending in 0 (Custom Darian) |
| 73 | + "IsLeap", MOD([Value], 2) = 1 || MOD([Value], 10) = 0, |
| 74 | + "SolsInYear", IF(MOD([Value], 2) = 1 || MOD([Value], 10) = 0, LeapYearSols, StandardYearSols) |
| 75 | + ) |
| 76 | + |
| 77 | + VAR CalendarMonthSols = |
| 78 | + SELECTCOLUMNS( |
| 79 | + GENERATE( |
| 80 | + IsLeapYear, |
| 81 | + GENERATESERIES(1, 24, 1) // 24 months |
| 82 | + ), |
| 83 | + [MartianYear], |
| 84 | + "MartianMonth", [Value], |
| 85 | + "SolsInMonth", |
| 86 | + IF( |
| 87 | + MOD([Value] - 1, 6) + 1 <= 5, |
| 88 | + 28, |
| 89 | + IF([Value] = 24 && [IsLeap], 28, 27) |
| 90 | + ) |
| 91 | + ) |
| 92 | + |
| 93 | + VAR CalendarDayBase = |
| 94 | + SELECTCOLUMNS( |
| 95 | + GENERATE( |
| 96 | + CalendarMonthSols, |
| 97 | + GENERATESERIES(1, [SolsInMonth], 1) |
| 98 | + ), |
| 99 | + [MartianYear], |
| 100 | + [MartianMonth], |
| 101 | + "SolOfMonth", [Value] |
| 102 | + ) |
| 103 | + |
| 104 | + VAR AddSolsSinceEpoch = |
| 105 | + ADDCOLUMNS( |
| 106 | + CalendarDayBase, |
| 107 | + "SolsSinceEpoch", |
| 108 | + VAR CurrentYear = [MartianYear] |
| 109 | + VAR CurrentMonth = [MartianMonth] |
| 110 | + VAR CurrentSol = [SolOfMonth] |
| 111 | + // Calculate total sols from all complete previous years |
| 112 | + VAR SolsFromPreviousYears = |
| 113 | + SUMX( |
| 114 | + FILTER(IsLeapYear, [MartianYear] < CurrentYear), |
| 115 | + [SolsInYear] |
| 116 | + ) |
| 117 | + // Calculate sols from complete previous months in current year |
| 118 | + VAR SolsFromPreviousMonths = |
| 119 | + SUMX( |
| 120 | + FILTER(CalendarMonthSols, |
| 121 | + [MartianYear] = CurrentYear && [MartianMonth] < CurrentMonth |
| 122 | + ), |
| 123 | + [SolsInMonth] |
| 124 | + ) |
| 125 | + RETURN SolsFromPreviousYears + SolsFromPreviousMonths + CurrentSol |
| 126 | + ) |
| 127 | + |
| 128 | + VAR Combine = |
| 129 | + ADDCOLUMNS( |
| 130 | + NATURALLEFTOUTERJOIN( |
| 131 | + AddSolsSinceEpoch, |
| 132 | + MonthNames |
| 133 | + ), |
| 134 | + "Date", [SolOfMonth] & "-" & [MonthName] & "-" & [MartianYear] |
| 135 | + ) |
| 136 | + RETURN |
| 137 | + Combine |
| 138 | + ``` |
| 139 | + |
| 140 | +Since the `[Date]` field is not a tradition date, we are not able to mark this as a date table. But we are still able to define our calendar. |
| 141 | + |
| 142 | + |
| 143 | + |
| 144 | +And we can create some fake weather data to test our calendar. |
| 145 | + |
| 146 | +??? demo "Weather Table" |
| 147 | + |
| 148 | + ```dax |
| 149 | + FactWeather = |
| 150 | + VAR BaseTemp = -60 // Average global surface temp on Mars in Celsius |
| 151 | + VAR TempRange = 25 // Typical daily temperature variation |
| 152 | + VAR MaxDailyVariance = 5 |
| 153 | + RETURN |
| 154 | + ADDCOLUMNS( |
| 155 | + SELECTCOLUMNS( |
| 156 | + Dates, |
| 157 | + "Date", [Date], |
| 158 | + "AvgTemp_C", BaseTemp + (RAND() * TempRange) |
| 159 | + ), |
| 160 | + "MinTemp_C", [AvgTemp_C] - (RAND() * MaxDailyVariance), |
| 161 | + "MaxTemp_C", [AvgTemp_C] + (RAND() * MaxDailyVariance) |
| 162 | + ) |
| 163 | + ``` |
| 164 | + |
| 165 | +Then we are able to use any of the regular time intelligence functions by specifying our calendar, like seeing what the weather was like last year. |
| 166 | + |
| 167 | +```dax |
| 168 | +Avg Temp Previous Year = |
| 169 | +CALCULATE( |
| 170 | + [Avg Temp], |
| 171 | + SAMEPERIODLASTYEAR( 'Darian') |
| 172 | +) |
| 173 | +``` |
| 174 | + |
| 175 | + |
| 176 | + |
| 177 | +## Conclusion |
| 178 | + |
| 179 | +Calendar-Based Time Intelligence opens up exciting possibilities beyond traditional Gregorian calendars, and final Martian's can use DAX's Time Intelligence functions. |
0 commit comments