Power Query number rounding issue

[Trebor84]

Hi, I have a list of accepted times and cancelled times in a table and I am trying to calculate the difference in minutes rounding up to the next minute.

ACCEPTED	Cancelled	Minutes Until Cancelled
01/01/2019 02:06	01/01/2019 07:06	300
01/01/2019 02:10	01/01/2019 07:09	300
01/01/2019 02:10	01/01/2019 07:46	336

In Excel the Minutes Until Cancelled column works as expected using the below formula:

=ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,"1:00")

I am trying to replacate this in Power Query but can't get the rounding correct using the formula:

=Number.RoundUp(Duration.TotalMinutes([Cancelled] - [ACCEPTED]),1) 

If I apply an additional step on the field using the built in Power Query Round Up function, I get this which is better but not sure why my original step isn't producing the same result?  Also this is still slightly wrong as the second value needs to be 300 minutes not 299.  Taking into account the seconds the value probably is 299 however I want to move this up to the next minute which the Excel formula achieves.

Any suggestions please?

[Vijay_A_Verma · Accepted solution]

1. =ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,"1:00") - "1:00" is a character, hence it will be treated as 0 in Excel. Hence =ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,0) will give the same result.

2. Now, below are results for [@Cancelled]-[@ACCEPTED] in Excel and when this is multiplied by 1440.

Below screenshot is from Excel

Now, you see the second result is 299.00000000023200. Excel's precision is 15 digits. Now, if you count 15 digits, the number would be 299.000000000232. Since Roundup is detecting 232 after decimal places, hence Roundup is making 299.00000000023200 to 300.
3. This behaviour is same in DAX also. Hence, DAX and Excel behaves the same.

4. See the below for results of Current-Accepted (This screenshot is from Power BI Desktop)

Now let me use Windows supplied calculator for second row values

DAX - 0.207638888889051*1440 : 299.0000000002334

PQ - 0.207638888888889*1440 : 299.0000000000002 

Now, in case of PQ when I count 15 digits, it comes out to be 299.000000000000. This is quivalent to 299, hence no effect due to Roundup. 

Hope, it clears up the issue for you. 

I will also suggest you to read Floating point Airtmetic issue of Excel - https://docs.microsoft.com/en-us/office/troubleshoot/excel/floating-point-arithmetic-inaccurate-resu... 

[collinsg · Accepted solution]

The second argument of Number.RoundUp is “digits” – the number of digits to round to. Replacing the 1 in your formula with a 0 will fix the problem.

Number.RoundUp( 
   Duration.TotalMinutes( [Cancelled]-[Accepted] ),
   0
) 

[Trebor84]

Thanks, yes that makes sense as to why Power Query displayed a different result.

I used a dax calculation in the end to achieve the desired result.

[Fowmy]

@Trebor84 

I am not sure If @MattAllington  meant the same as I tried this way

Number.RoundUp(Duration.TotalSeconds([Cancelled]-[ACCEPTED])/60)

[Trebor84]

Hi, that works apart from the rounding up.

In Excel the second example gives me 300 minutes which is what I am looking for using: =ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,"1:00")

[Vijay_A_Verma · Accepted solution]

1. =ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,"1:00") - "1:00" is a character, hence it will be treated as 0 in Excel. Hence =ROUNDUP(([@Cancelled]-[@ACCEPTED])*1440,0) will give the same result.

2. Now, below are results for [@Cancelled]-[@ACCEPTED] in Excel and when this is multiplied by 1440.

Below screenshot is from Excel

Now, you see the second result is 299.00000000023200. Excel's precision is 15 digits. Now, if you count 15 digits, the number would be 299.000000000232. Since Roundup is detecting 232 after decimal places, hence Roundup is making 299.00000000023200 to 300.
3. This behaviour is same in DAX also. Hence, DAX and Excel behaves the same.

4. See the below for results of Current-Accepted (This screenshot is from Power BI Desktop)

Now let me use Windows supplied calculator for second row values

DAX - 0.207638888889051*1440 : 299.0000000002334

PQ - 0.207638888888889*1440 : 299.0000000000002 

Now, in case of PQ when I count 15 digits, it comes out to be 299.000000000000. This is quivalent to 299, hence no effect due to Roundup. 

Hope, it clears up the issue for you. 

I will also suggest you to read Floating point Airtmetic issue of Excel - https://docs.microsoft.com/en-us/office/troubleshoot/excel/floating-point-arithmetic-inaccurate-resu... 

[MattAllington]

I think you will find that TotalMinutes ignores the seconds. Try doing the calc in seconds, then converting to mins and rounding then. Just a suggestion. 

[Trebor84]

Still having trouble with this, so I have tried seconds.

=Duration.TotalSeconds([Cancelled]-[ACCEPTED])

I have then converted to what I think should be minutes.

=Duration.TotalMinutes([Seconds])

Expression.Error: We cannot convert the value 17953 to type Duration.
Details:
Value=17953
Type=[Type]